At iRobot, moving way beyond the Roomba
Road Trip 2010: CNET reporter Daniel Terdiman stops in on the maker of the hit robotic vacuum cleaner and learns what the company has in store for the future of military and home robotics.
BEDFORD, Mass.--I have seen the future of military robotics, and it is autonomy.
I've come here to visit the headquarters of iRobot, the company probably best known for its famous Roomba vacuum cleaners. But while it has sold more than 5 million of those cute household devices, it has also developed a reputation as one of the world's leaders in designing battlefield-ready robots capable of things like detecting and extracting explosive devices, search-and-rescue, and much more.
And though iRobot could probably rest on its laurels, and keep making profits for the foreseeable future with the Roomba and military-grade robots like the PackBot or the SUGV--the small unmanned ground vehicle--it is at the same time looking to set the pace for the robots that will man--so to speak--the battlefields of the future.
Naturally, iRobot is also looking for ways to advance its home consumer line of robots, and in that realm, it believes its path to down-the-road profits will depend heavily on health care and care giving. But I'll get to that later.
The techie admiral
I've come to iRobot as part of my Road Trip 2010 journey through the American Northeast, in search of a clear picture of what one of the few successful, large-scale robotics companies thinks the future looks like. Of course, I've also come to see a bit of the company's history, and during my visit, I got a really nice look at both.
One of my stops during the visit was with a gentleman who definitely isn't from Massachusetts--Joseph Dyer, iRobot's president of government and industrial robots division. Dyer, a genial Southerner with a discernible if not thick accent, is currently in his "second career" after years spent in the Navy, where he retired as a vice admiral. But thanks to his focus in the military on tech and how it can help the military achieve its goals, he became known as the "techie admiral," I'm told.
Dyer doesn't mince words about why he came to iRobot: "I wanted to find the 2000s equivalent of Apple in the '80s," he tells me as we sit down for an information-packed 15-minute talk.
With the PackBot, the SUGV and other devices that have been in the field for years, iRobot has already made a difference to a lot of people, and Dyer said that is established all the time with the arrival of postcards from soldiers in Iraq and Afghanistan who, based on those devices' ability to sniff out threats like improvised explosive devices (IEDs), frequently write things along the lines of "You saved lives today."
"If you're looking for gratification for coming to work," Dyer says, "it doesn't get any better."
Admittedly biased, Dyer said that he sees robotics as the next big American field of innovation after a century of success in areas like agriculture, automobiles, information technology, and more.
But how can iRobot and other companies in the industry take a field in which these small devices are already capable of helping to find and destroy explosives, do search-and-rescue missions and much more? To Dyer, the answer is in solving one of the key bottlenecks that to date has limited the full range of what's possible with robotics in the battlefield.
Until now, he explained, there has always been a one-to-one relationship between the robots and the soldiers who operate them in the field. But that ratio is about to change dramatically, Dyer predicted. "The one word answer to why," he said, "is autonomy."
He likened today's robots to the state of fighter planes 35 years ago, when limitations in electronics and other technologies meant that a pilot had to devote 80 percent of his or her time to attending to the plane's airframe, engine, and navigation, leaving just 20 percent to their mission. But over time, thanks to significant advances in fighter technology, those numbers are almost entirely reversed, Dyer said.
Similarly, with today's telepresence robots, about 80 percent of the operator's time is spent focusing on the technology, while just about 20 percent is available to focus on the mission. But thanks to things like Moore's Law, he predicted, it will take far less than 35 years to reverse that ratio, meaning that within a few years, the technology will exist to make it possible for a human operator in the field to focus almost entirely on the task at hand and not have to worry much about the practicalities of getting their robot to do what they want.
Indeed, Dyer said that he expects the first major shift in this area to take place later this year, with the delivery of the first truly autonomous robots to soldiers in the field.
To be sure, these robots won't be that much more advanced than what is out there today, but Dyer said they will feature autonomous communication technologies known as "retro-traverse" that will allow them, for the first time, to navigate out of trouble if they lose communications connection with their human operators.
Today, he said, if a robot loses communications, soldiers are required to take the time and the personal risk to go find it, something that is entirely counter-productive considering that the robot is meant to obviate soldiers putting themselves in harm's way.
But the new generation of robots should be smart enough, Dyer explained, to figure out that if they lose communications, they should back-track far enough to where they can once again pick up the signal.
At the same time, the new robots will also feature cruise control, which will mean that they can automatically maintain course and speed. That may not sound like much, but to a soldier who currently has to take care of those elements manually, it could be a major shift forward.
Looking further down the line, Dyer said that by 2015 or so, soldiers should be getting their hands for the first time on so-called "robot wingmen."
The idea here, he explained, is for a robot to be able to autonomously take on task assignments for things like getting through closed doors. The robot won't have full artificial intelligence, but based on a programmed mission profile, it should be designed to attack a task and generally figure out how to complete it. Even better, if it encounters problems, it won't simply stop working, it will be smart enough to ask for help.
That might mean that if a robot is assigned to get through a door "peacefully," but finds that the door is locked, it can inquire--by voice, or even by text message--how it should proceed.
And, human operators may also be able to give instructions to robots simply by making hand gestures--something that could free up a lot of attention that is currently spent manually handling controllers.
Finally, Dyer said, the more autonomous robots are deployed in the field, the more that will free up already-taxed communications networks since it requires a great deal of bandwidth to maintain connectivity between soldiers and their robots.
According to Chris Jones, iRobot's head of research, a significant percentage of the forward-looking work that the company does, especially on the military side, is funded by what is known as "6-1 money." This funding comes from DARPA--the Defense Advanced Research Projects Agency--and is essentially defined as very early stage, or "basic" research. In other words, work done to figure out what's possible, not necessarily what to do with it.
With that in mind, iRobot's research division focuses on four main areas: Autonomy--which Dyer obviously sees as a big part of the company's future; Human-Robot Interaction, which will, in part, utilize autonomy to help users more efficiently control their robots; Collaboration--which could mean single users can simultaneously control multiple robots; and New platforms.
As an example of the last one, Jones demonstrated a new project based on "6-1 money:" the so-called "Chemical Robot." This is a project with no specific known application, and looks nothing like any robot you've seen before. In fact, it looks more like a blob of soft plastic.
But what it is is a collection of soft material that can, on demand, be expanded or retracted, all with the idea that it can be rolled somewhere and then squeezed under or through very small spaces. It may be the size of a softball in its full form, but could theoretically get through a hole the size of a quarter, Jones predicted.
While there are no specific tasks in this device's future, it's thought that search-and-rescue is a major possible application.
Generally, there are no known timeframes for the productization of technologies like this, Jones added, but said that the point is first to prove they are viable, and then what to do with them.
Similarly, as iRobot works on new-style versions of the PackBot that are smaller and, thanks to better sensors and technology, more autonomous, the answer to when they are deployed in the field may have more to do with military culture than technology.
My last stop of the day was with iRobot founder and CEO Angle. For some time, we talked about his company's history, and how he came to develop the Roomba and its successor, the floor-cleaning Scooba.
But Angle also had a forward-looking message. To him, iRobot's commercial future is centered on health care and care-giving. He explained that over time, as humans live longer, the number of people available to take care of the elderly--family members and friends, mainly--is shrinking due to major cultural and demographic shifts.
And since most elderly people want to stay away from nursing homes, that means there's a huge opportunity for a company like iRobot--and its competitors--to develop robots that can fill the resulting gap.
And that means that if iRobot or other companies can come up with home-based robots that can do things like proactively remind elderly people to take their medications, or to get out of bed, go to the bathroom, or any of many other possible tasks, it could make a huge difference in people's lives.
That might not be intuitive to the average robotics fan, but it's crucial to the industry, Angle said.
"The industry as a whole has always had a weakness in that robots are so cool that people have always been excited" by their futuristic possibilities, Angle said. "What the industry needs is people solving real problems."
For the next few weeks, Geek Gestalt will be on Road Trip 2010. After driving more than 18,000 miles in the Rocky Mountains, the Pacific Northwest, the Southwest and the Southeast over the last four years, I'll be looking for the best in technology, science, military, nature, aviation and more throughout the American Northeast. If you have a suggestion for someplace to visit, drop me a line. In the meantime, you can follow my progress on Twitter @GreeterDan and @RoadTrip and find the project on Facebook. And you can also test your knowledge of the U.S. and try to win a prize in the Road Trip Picture of the Day challenge.