Colin Angle’s presentation comes across as an unexpected combination of a TED talk on the future of robotics and a paean to the virtues of scrubbing your kitchen floor more regularly.
That’s because Angle, the CEO of iRobot – maker of the Roomba robot vacuum cleaner – knows that while talk of artificial intelligence is fascinating to some, his devices will only sell if he can convince the average consumer that bots will do a better job of cleaning up than they could themselves.
“The industry has been traditionally been full of companies that are so excited by the technology, by walking android-like machines, that they forget why they are building robots in the first place,” he says.
The history of iRobot follows a similar theme, with exotic but small projects gradually giving way to the perhaps more mundane, but much more mass market, adoption of robots.
The company has been building robots for close to a quarter of a century, starting out with developing the most boys-own bots imaginable, ones designed for lunar exploration. This was followed by robots used in bomb disposal and other military settings; then the company moved into the consumer space with toys and finally the robot vacuum cleaners that have brought iRobot mainstream success.
The company has now sold 10 million of its Roomba autonomous vacuum cleaners – squat, over-grown-hockey-puck-bots which look about as far from anthropomorphic androids as possible.It’s probably the opposite to how you’d imagine a robotics company’s products would evolve, to start with space exploration and end up in vacuuming, but for Angle it’s thinking like that which is precisely why robots have failed to break through sooner.
“Trying to think of robots as mimicking humans is a mistake. We should not – as the robot industry tried to – create artificial people that look like artificial people unless we’re doing entertainment,” he says.
Around 90 percent of iRobot’s business is these domestic robots and even if for some paying £700 for a robot vacuum cleaner or £450 for a robot to mop your floor might seem extravagant, it’s a market that is growing at 30 percent per year, which for Angle vindicates the company’s focus on the domestic over the gee-whiz.
Instead of droids that look like us, Angle would prefer his bots to be invisible. “The perfect vacuum cleaning robot is invisible, you don’t talk to it, it doesn’t flash with lights, you never see it, you never touch it. That’s what we’re shooting for,” he says.
Meet the ‘throwbot’
While the vacuum cleaners are the best known of iRobot’s devices, the company’s current portfolio stretches from domestic robots through to military devices and now business robots, which I saw at a launch event in Munich. For example around 5,000 of the company’s PackBot military robots have been deployed into warzones to be used as part of bomb disposal, and are able to lift 320lb and climb over very difficult terrain.
“In some ways you might think you are looking at a very exotic impressive machine but I would tell you what you are truly looking at is the world’s first practical human,” Angle said. The point he is making: just because a robot takes on a job that a human would otherwise do, that doesn’t mean that it has to look like a human.
In contrast to the large scale PackBot, the company’s ‘FirstLook’ is a shoe-box sized robot on tracks, which soldiers can use to explore areas too small or too dangerous for humans. It’s also known as the ‘throwbot’ because it’s light enough to be thrown onto a roof or (double glazing permitting) through a window to get a better view from its multiple cameras.
Throwing a robot worth tens of thousands of dollars across the room is a new experience for me, but the iRobot staff are more far worried about me smashing a glass panel in a nearby staircase than they are about me damaging the bot. When someone inadvertently steers the diminutive robot under a raised platform the staff simply switch on its infra-red camera to help them steer it back out of the darkness.
Despite their differences in scale and capabilities, many of the machines in iRobot’s lineup share common elements or heritage; the ‘iAdapt’ algorithm the Roomba uses to make sure it has thoroughly cleaned a floor was originally developed for the US Department of Defence for mine hunting. They’re also united in their focus on utility, which Angle says has been key to the company since its inception in Boston in 1990.
“We did have this notion that we had to be solving practical problems and the way we were going to create this industry was robots that delivered more value than they cost to build,” he said.
The Roomba vacuum cleaner came out of a two-engineer $15,000 proof-of-concept project which sold 70,000 robots in the first three months, with demand so great that retailers were willing to pay for the machines to be flown in from China.
Vacuuming might seem to be a rather dowdy usage of artificial intelligence and engineering smarts, but Angle sees it differently. “I think it says the industry is growing up a bit because if you want to be a big major player in the world then you have to be solving a big major problem,” he says.
And after years of failing to live up to the science fiction ideal, robotics has suddenly become an incredibly hot area.
Robots – the next big thing
Over the last few months Google has bought some of the biggest robotics companies including Boston Dynamics (which created the ‘Big Dog’ rough terrain robot that can carry heavy loads) Meka Robotics and Redwood Robotics, all of which will be part of Google’s robot strategy headed (fittingly) by Android creator Andy Rubin. Google’s work on driverless cars also feeds into the logistics jigsaw; think convoys of driverless trucks.
Similarly, Amazon bought warehouse automation company Kiva back in 2012 and has showcased its plans to use drones for deliveries, a service it calls Prime Air.
Robots are intriguing for these big tech companies for a couple of reasons. One major factor is simply the operational efficiency that robots can offer – in this case the ability to automate tasks, from picking products in a warehouse to cheaper delivery; a new variation on the dull, dirty, dangerous type jobs that robots have been used in for many years. Anything that can get products picked and shipped faster can give one e-commerce retailer the cost saving it needs to leap ahead of the competition.
But at a deeper level these tech giants are also intrigued because robots bridge the physical and digital worlds. The sensors on a robot can provide much more information about their human owners than their web history or smartphone data. And as the smartphone market reaches saturation, robots become an obvious next step for the tech giants keen to gather as much information about the real world as possible.
This combination of sensors, robotics and smarter algorithms are all elements of the internet of things, although when it comes to robots and sensors in the home, we’re really talking about the internet of us. For example, researchers have already demonstrated just how much you can infer from some relatively basic sensor data such as power consumption, CO2 levels and humidity.
Pretty much every major tech company wants to provide that gateway to the connected home whether that’s through entertainment; you can argue the Xbox One is part of Microsoft’s strategy here, or through sensors; Google’s acquisition of Nest implies some intriguing futures; or through robots.Angle seems relaxed about which company wins the battle to own the software gateway to the home, saying “we’ll let them fight it out,” but wants iRobot’s devices to be the hardware, the physical presence within the home, and foresees a “menagerie of robots” cooking and cleaning, monitoring and maintaining.
These robots won’t look like humans – they don’t need to. A washing machine and dishwasher both do jobs originally done by humans, but neither have arms and legs. But Angle thinks there will soon be one android-style robot in the home: “There is a robot that maybe brings it all together that is focused on interacting with you, this human interface robot, this butler robot.”
You can see the first steps towards this in one of iRobot’s latest models. The Ava 500 is the only remotely human-looking robot that the company has, made up of a column topped with a large screen; a roving video-conference robot which can be programmed to meet you at your desk or find its own way to a meeting room. Or glide around the office to come and find you at your desk if you try to skip a meeting (like a Dalek, it can’t climb stairs. Yet.)
While the robot is currently sold to the business market, Angle sees this deployment as a proving ground for its eventual move in to the home.The idea is to create a more physical presence for the person at the other end; the screen can be lowered to simulate sitting down, or turn to address other people at a meeting.
Talking to someone via the robot – here and yet not here – is a disconcerting experience because giving them a physical, gliding mobile presence through the robot really does alter the dynamic in a way that a static webcam chat on a laptop never would. It even puts the language we use under strain, plasticising some of the basics of discourse; when I’m talking to one of the iRobot execs who has stayed back at the Boston HQ I ask: “What’s it like for you to be here?” even though they aren’t ‘here’ and the ‘you’ isn’t them.
Angle sees something like this as the basis – a decade down the line perhaps – of a general purpose robot companion that would manage the other robots in the home, as well as allowing doctors to carry out remote diagnosis by taking control of it, or friends or family to ‘visit’.
It’s not the only vision of the future and many might be reluctant to share their homes with a robot, preferring to build that sort of intelligence into the building itself; like Siri embedded in the walls of your home. Angle argues that physical robots give you a better chance at privacy than an all-knowing smart house because you can at least slam a door on a robot.
Still not everyone is convinced that we’re about to see the dawn of domestic robots. Making the leap from vacuuming to more general application of robots could still be a long way off, says Martin Smith, Professor of Robotics at Middlesex University: “The thing about robot floor cleaners is that is the easiest task a robot could do. If you want a robot to do anything more complex you’re really stuck. If you wanted it to iron a shirt that is terribly difficult, even walking up stairs is difficult for a robot unless it has tracks.”
He adds: “There’s a kind of romantic notion that we are going to build a robot human in my lifetime that’s going to be useful and economically viable. I suspect it won’t happen.”
Robots in an ageing society
But it could be that in the longer term robots are the only way of defusing the demographic time-bombs that are ticking across much of the developed world.
Governments are already finding it harder to pay for the care needed by their ageing populations – and the staff to do those jobs too.”We need to find technology that will allow people to live independently in their own homes which happily is what they want to do anyway for as long as possible or else society is going to face a decrease in its standard of living,” Angle warns.
For many of gen X-ers in 30 years time it may become a choice between living in our own homes with a robot, or moving into some sort of assisted housing as we get old. At which point that robot stops looking like a Dalek and a lot more like your best friend.
Of course there are plenty of other companies in the area, but Angle remains convinced that robots are the answer. “If you are rooting for iRobot great, if not root for someone else. But root for robots to come in and make our ability to live independently greater and greater over time – because we need it.”
TechRepublic travelled to Munich to meet Colin Angle as a guest of iRobot.