"I have seen the future and it sucks. It also sweeps and removes dust and pieces of dirt," wrote Mashable Editor Lance Ulanoff in 2002. "Its name is Roomba....For those of you who are thinking about taking this development lightly, don't. While we were all going gaga over Dean Kamen's It (Ginger, Segway Scooter, whatever), iRobot CEO Colin Angle and his company were quietly developing a product that may really change the world."
In the late 90s and early 2000s, domestic robots were a booming business--in pop culture, at any rate. Steven Spielberg's A.I. presented a dystopian look at an android-filled future; the Star Wars prequels brought C-3PO and R2-D2 back to the big screen for a new generation of kids; even Bicentennial Man gave life to an imaginative (if hopelessly sappy) future of fleshy, feeling robots--all fit for service for the neo-future domicile. But Robbies and Rosies and Marvins and Terminators have been key players in science fiction since the 1950s. The difference was, when September 2002 rolled around, you could actually buy one for your own home.
It couldn't walk or talk, but the Roomba was at last a practical domestic bot, and it cost $200. People took notice. Time magazine called it one of the best inventions of 2002. And everyone agreed that home vacuuming totally sucks--so why not let a robot do it?
A decade later, Massachusetts-based iRobot has released three distinct generations of Roombas and sold over eight million sweeping, vacuuming homebots. They weren't actually the first to market with a floor-cleaning automaton--Electrolux released the Trilobite in 2001--but the Roomba was the first to catch the public eye (and it cost about $1500 less than the pricey Trilobite). It's been getting better ever since.
"People have a vision of this multi-function, multi-purpose robot that can do everything except fly," says Maurice Leacock, a Senior Product Manager at iRobot. "Well, I guess Rosie can fly too. [Roomba] was the first way we could test the hypothesis people had that this was going to happen."
In light of the Roomba's 10th anniversary, I chatted with Leacock about Roomba's origins and the challenges that still stand between the robotics industry and a full-on home invasion. Despite Roomba's success, the home robotics revolution that seemed to be dawning at the beginning of the 21st century still isn't here. It's time to adjust our expectations.
Roomba and the Robot Promise
Earlier this year, iRobot CEO Colin Angle talked to The Wall Street Journal about the challenges facing commercial robotics companies. Rather than talking tech or pricing, he made an interesting point: Our perception of robots can stifle innovation (or acceptance, at the very least).
"In 1962 The Jetsons showed us Rosie and we all became completely convinced what a robot is and what it should do. It's 50 years later and there is no Rosie, and I think that so many of us became skeptical that robots were ever going to be part of our lives that it was almost unconscionable that "oh wait, I can go buy a robot vacuum today and it will actually work." So skepticism has been the biggest barrier to growth in the industry."
Anthropomorphic robots like Honda's ASIMO fit the common perception of what a robot should look like, but shooting for humanoid designs can be problematic. It's harder and more expensive to make a walking robot than a rolling one. Too human, and you hit an uncanny valley that weirds people out (Japan's excelled at this for years).
iRobot brushed up against the valley issue with My Real Baby, a toy they brought to market in 1999, three years before the Roomba. Based on two emotionally responsive robot prototypes named IT and BIT, My Real Baby was designed to have facial expressions and enough sensors and motors beneath its rubber skin to approximate the real thing.
"We tried to be on the positive side of that uncanny valley," says Leacock. My Real Baby would "learn" over the course of 18 months, changing its sounds and emotional responses. "That was probably way ahead of its time and probably still would be ahead of its time--there's a very fine line between creepy and delightful." In other words: The idea was great, but the technology wasn't quite ready.
The uncanny valley wasn't a problem for Roomba, which began development at iRobot about a year before My Real Baby launched in partnership with Hasbro.
"Over two to three years it went through three to four really prototypey phases," says Leacock. "Products you wouldn't want in the home, but something you'd want to take to a manufacturing partner."
My Real Baby provided a valuable opportunity to build relationships with manufacturers in Asia. After three years of designing, Roomba was ready to graduate to manufactured consumer product. In a way, Roomba's plastic disc body was a way to lower, or at least change, consumer expectations for a robot you could buy for the home. While My Real Baby got a bit lost in the uncanny valley, Roomba looked about as utilitarian and un-robot-like as possible.
That's a core part of Roomba's design: Single-minded utility. "If you want something that can clean your floor all the time, every time, that's a goal we can meet," says Leacock. We're still a long way off from robots that can walk around the house, cook and clean, which means Roomba will always be fighting against that ingrained pop culture image of what, exactly, a robot should look like.
But that raises another question: Would we actually want a life-size domestic service bot if the technology existed? "Duh" is the snap response, yet every Internet article about robots is guaranteed to pull in a few Skynet and Terminator responses. Isaac Asimov's writings about the Three Laws of Robotics (which dictated that robots would not harm humans, and so on) often as not involved robots bending or breaking those laws under unique circumstances. We yearn for more intelligent robots, but we're a bit afraid of them, too.
"Maybe we can reach a point where all the robots can talk to each other," says Leacock, envisioning a home full of single-use robots dedicated to everyday duties. With a robot like Rosie "we hit that uncanny valley again...I'm not sure how much I want my robot knowing about me," he adds with a laugh. "People will become more comfortable with machines--which are what robots are--being parts of their lives."
If a Rosie in every home is impractical, what should we expect to be the Next Big Thing in robotics? Odds are good the robotic revolution won't start in the home--high-end Roombas cost up to $700, and that's nothing compared to industrial and military machines. iRobot's PackBot, which was used for reconnaissance and bomb disposal in Iraq and Afghanistan and was the first to scope out Japan's damaged Fukushima reactor in 2011, sold to the military for $60,000 per bot. UAV Predator drones cost a cool $4 million per flyer.
More advanced needs, more money. More money, more innovation.
The Next Big Thing could be a robot like Baxter, which costs about $22,000 and offers a new degree of versatility for a factory robot. Instead of replacing a machine that can laser etch circuit boards with inhuman precision or stamp metal with thousands of pounds of force, Baxter can perform human tasks with two flexible arms and can easily be programmed via a touchscreen that doubles as a face. He doesn't do things we can't do--he does things we don't want to do.
Rethink Robotics, Baxter's creator, was founded by roboticist Rodney Brooks (who also co-founded iRobot in 1990). iRobot has its own robot in the works that could push the industry forward. Like Baxter, iRobot's Ava is targeted at businesses, rather than homes, and will likely cost tens of thousands of dollars. If Ava is successful, that price will presumably come down over the years and those innovations will make it into robots the average person can afford.
As Roomba has proven, though, it takes years and years to iterate on a single task when you're designing a robot that the average person can afford.
Roomba Grows Up
At a glance, the Roomba's barely changed over the past 10 years. It still looks like a landlocked frisbee. It still rolls around and cleans your house. But there have been changes. In 2004 iRobot eliminated the buttons on the Roomba for small, medium and large rooms--the robot could determine that for itself thanks to a better microprocessor. Pedometers added to the wheels tracked how far the Roomba had rolled. The 2011 models saw improved acoustic sensors for detecting dirty areas of the floor and a new bin design for pulling in more dirt.
The same advancements that have benefited smartphones and laptops are important for the Roomba, too. But new technology isn't cheap: The entry-level Roomba now runs $350, 150 bucks more than the original cost a decade ago, while the top-end model costs $700.
"The fact that batteries can last twice as long as just a few years ago is a big deal," says Leacock. "Advancement of microprocessor technology, robot technology...The fact that cameras and cellphone technology are becoming cheaper and cheaper everyday allows you to recognize different obstacles. [We'll be] doing things in the coming years we could only imagine a couple years ago."
True to their consumer philosophy, iRobot hasn't added any features to Roomba to expand its singular purpose; instead, they've launched additional household bots to take care of the mundane tasks we all hate. The Scooba mops floors, the Verro cleans pools, and the Looj tears through gutter grime.
On September 17, 2012, iRobot announced a $74 million acquisition of Evolution Robotics, makers of a hardwood floor cleaning bot called Mint. While Evolution's technology will likely feed into iRobot's existing machines (and vice versa), the purchase reinforces Leacock's point that the Roomba is only a small part of iRobot's business. It's just the part that caught on.
Roomba excited the world because it was the first indication that The Jetsons, or at least some vision of a robot-filled 21st century, was within reach. We may be waiting decades for the likes of Honda's Asimo to bring affordable multifunction robotics into the home, but that robot-filled future is already sneaking up on us.
A new Honda lawnmower bot will cost about $3000 in 2013, and it's not the first. In a few years we could have robots for seeking out foul odors around the house. A few more, and it's not unreasonable to think that owning robots to clean our gutters, cut our grass and polish our floors will be relatively affordable.
In another 10 years, Roomba will likely still be there sucking up dirt. It'll be a little faster, a little smarter, and hopefully will be one of many robots keeping our houses clean. It won't talk, but it'll probably still bump into furniture every so often, just to remind you that it's still there.