Where's My Robot Maid?
|
|
CREDIT: Hanna-Barbera Productions/Warner Bros. Television |
When computer science professor Siddhartha Srinivasa's Personal Robotics Lab at Carnegie Mellon had an all-day party for hundreds of people, guests could order their drinks via their smart phone or in person. The caterer, HERB, was methodical in his drink delivery and maniacal about picking up empty bottles for recycling. Srinivasa knew HERB was perfect for the job; he was programmed that way.
HERB, short for Home Exploring Robot Butler, is the promise that the next generation of robotics holds for anyone who has ever dreamed of outsourcing chores to a machine. The all-purpose robot has long been touted in science fiction and cartoons, but robotics has yet to leap out of factory floors and into homes in any significant way. Newer technology and better algorithms are moving the field forward significantly, although challenges remain with sensing, dexterity and processing.
"I am personally a big fan of a general purpose robot," Srinivasa told InnovationNewsDaily. "You can have more and more of these specific purpose devices, like the Roomba, but eventually you just run out of space. There is a need for more integration."
Robots like HERB, who is essentially two arms and a computer on a Segway base, will need complex algorithms. But they'll also require and the ability to manipulate their limbs with the finesse of humans to open doors or mop the floor -- or pick up that coffee mug off the kitchen table without crushing it. Ideally, the machine will be able to learn new skills along the way. "My dream is for HERB to build an Ikea shelf," said Srinivasa. The robot is already learning to drill holes in preparation.
Various researchers are working on different incarnations of the robot maid , and it's unclear what form the at-home robot will take. Honda's ASIMO, which is a humanoid, took more than two decades of research to be able to walk and move similar to a human.
Replicating human movement is just one challenge. Another problem is learning new tasks in the real world. HERB, for instance, could take a drink order for a specific set of drinks, go get it and then find that person in a crowded room but he can't just look at a bottle of gin and know what to do with it if someone wants a dry martini. A robot at the University of California, Berkley can very slowly fold towels with great care, but it doesn't know what to do with a sock.
Optimists like Srinivasa say that better algorithms and better hardware are being developed every day, and it's impossible to know what breakthroughs will catapult the entire field forward. He guesses that within 15 years we might see the first incarnations of robots that can help with basic tasks, especially for at-home healthcare or in a hotel.
"The technology is getting better by leaps and bounds," said Srinivasa. "We want to shoot for the moon. Our research goals must be so absurd that they inspire us."
Follow InnovationNewsDaily on Twitter @News_Innovation, or on Facebook.
