The news that iRobot and Taser teamed up to unveil a “Taserbot” at a conference in Chicago this week made me wonder: where is commercial robotics around here really headed? Don’t worry, I’m not about to jump on the “Terminator” and “Robocop” bandwagon. I’m not here to complain about the militarization of machines. I’m just looking for some diversity.
Sure, iRobot has been a success by many measures. The Burlington, MA-based powerhouse has sold more than 2 million Roombas and has seen its military PackBot deployed by the hundreds in Iraq and Afghanistan. With the popularity of robo vacuum cleaners, it has gotten to the point where robotics researchers and industry types often don’t even call the commercial ones “robots” anymore—they call them “appliances.”
But perhaps iRobot’s approach isn’t the only way to make robots mainstream. That general feeling can be summed up by what Chris Atkeson, a roboticist at Carnegie Mellon University in Pittsburgh, told me not long after Roomba made its initial splash in 2002. “Robotics has been a failure for some time,” he said. “Too bad it succeeded by being stupid.”
There are other paths to commercialization. According to the Massachusetts Technology Leadership Council, some 150 companies in the greater Boston area do robots in one form or another. Some are industrial giants such as Foster-Miller in Waltham, but the majority are smaller startups focused on a niche application or two. Now they’re starting to step out and stake their claims to the market.
Take Boston-based Myomo, a privately held MIT spinoff that just announced this week that it has received FDA approval to market its first product. Led by cofounders Kailas Narendran and John McBean, Myomo has developed an advanced upper-arm brace that senses electrical signals from nearby muscles and activates a motor mechanism to help stroke patients control their elbow movements. There is a sizable market for this: stroke affects some 5.7 million people in the U.S. alone, and only 5 percent of survivors recover full use of their arms. The medical device company has recently completed research studies on the technology at Spaulding Rehabilitation Hospital in Boston, with promising results.
Some in the know say that within a decade, robots will provide even more advanced kinds of help. “A general-purpose robot to do a bunch of different things probably will become the killer app,” says Jonathan Klein, director of marketing for Vecna, a Maryland-based tech company that does robotics R&D at its Cambridge facility. “It will bring you not just your beer, but also your dinner, take you to places, help you get in and out of bed, read you the news, and even play computer games with you.”
That might be a stretch, but Vecna researchers have made some surprising progress. Take their “BEAR” (Battlefield Extraction-Assist Robot) prototype (pictured above), which is being developed with funding from the U.S. military. This robot, which gets around using a novel combination of wheels and treads, is designed to carry wounded soldiers from the battlefield to safety and could be ready for field testing within five years. This machine needs far more strength and skill than a Roomba or PackBot; the task of transporting a prone soldier is worlds more difficult than any current bot could safely be expected to achieve.
Where it gets interesting for consumers is that Vecna is starting to design a prototype for elder-care and health monitoring in homes, which Klein says it hopes to sell for a few thousand dollars (or more, depending on its features) within a decade or two. This home robot will include state-of-the-art navigation software and speech and social interfaces, to go along with its considerable mechanical know-how. It will also need to operate safely with little or no human supervision.
To see what’s coming further down the road, poke around in a few local academic labs, especially those with ties to industry. Domo, the upper-body bot from iRobot co-founder Rod Brooks‘s group at MIT, symbolizes the next big frontier in commercial robotics: dexterity. Born in the Computer Science and Artificial Intelligence Lab (CSAIL), Domo has a torso, two arms, two hands, and stereo cameras for vision. Using its advanced hand-eye coordination, it is learning to manipulate hand tools such as screwdrivers. The eventual goal of the project is to achieve the manual dexterity of a six-year-old child—a deep problem that could take 10 to 20 years to crack.
If and when such a level of dexterity is achieved, the commercial applications could be astounding: pretty soon you could have robotic repairmen, waitstaff, and personal helpers. To this end, an independent but related project on manual dexterity is being led by professor Rod Grupen at the University of Massachusetts in Amherst. His robot, “Dexter,” uses robotic arms and hands from Barrett Technology in Cambridge to gain experience picking up various objects, and to learn which grips to use for specific tasks.
How might these projects play out in the marketplace? Although most companies won’t pursue far-out research in areas like dexterity, they need to be thinking creatively more than ever. “The market for consumer robots hasn’t taken off,” says Klein, who was formerly a vice president at iRobot.
The Roomba, for all its success, has reportedly only made it into 1 to 2 percent of households in the U.S. so far—hardly a revolution in robotics appliances. Local experts aren’t concerned overall, though. “Toy robots and vacuum cleaner robots are the form factors that are breaking through to the consumers,” says Russ Tedrake, a professor at CSAIL who studies robot locomotion and collaborates with Boston Dynamics, a Cambridge-based robotics firm. “The next great robotic appliance will be some other small, niche application that most of us didn’t even think of.”
