The future of robotics lies in… arranging potted plants? Surprising, but it just might be true. Tonight at the MIT Enterprise Forum of Cambridge, a small company called Q Robotics, based in Groton, MA, gave its first public presentation on the technology it has been developing for the past year and a half. Founded by iRobot veterans Joe Jones and Paul Sandin, co-inventors of the Roomba vacuuming robot, Q Robotics’ project has been shrouded in secrecy. Now the veil has been lifted, and what’s underneath is more interesting than it might seem at first glance.
I caught up with Jones, the CTO and 24-year robotics veteran, by phone earlier today. The story of Q Robotics began the day before Thanksgiving in 2006, when Jones and Sandin left iRobot to start a new robot company. “We wanted to invent something, but we didn’t know what,” says Jones. “We wanted to build more practical robots that people could afford, that served actual needs, and that we could build in the near-term… Floor-cleaning robots are wonderful and make a good business, but what do you do next?”
In February 2007, they got an idea at a trade show in Boston called New England Grows, for people who grow ornamental plants—which are bought by everyone from landscapers and homeowners to dentists’ offices. It turns out there are several thousand growers of potted plants scattered around the country, and they grow the plants by the millions on vast fields. Moreover there’s a widespread problem in the industry, and it has to do with the spacing between pots. As the plants mature, the pots need to be moved around and adjusted to keep plants from growing into one another. Growers spend tens of millions of dollars a year on manual labor, just to monitor the pots’ positions and space them correctly in the fields.
Enter Q Robotics. Jones and Co. are building a team of small robots to roll around in fields, pick up potted plants, and space them appropriately. Each robot is roughly cylindrical, 20 inches in diameter and 15 to 18 inches tall, with two big wheels. The latest model (see photo) has a forklift-like mechanism with which to pick up pots. Each unassuming bot also has infrared range sensors to detect positions of objects and an optical sensor to tell whether it has properly grabbed a pot.
“It’s not glamourous, but there is a clear market need for this,” says Jones. And there are plenty of follow-on opportunities in agriculture that would use similar technology, he says, such as soil monitoring, sorting, and loading and unloading goods.
And the broader significance to robotics? “We wanted to make robots that were one technology step harder than what we did before,” says Jones. “Earlier robots like Roomba functioned by rolling around and not getting stuck. These are different, they actually manipulate the environment.” To Jones’s knowledge, that kind of manipulation—performed out in the “unstructured” real world and not on a factory or warehouse floor, as in the case of Woburn, MA-based Kiva Systems—has not been done successfully before. In other words, it might be one small step for a Roomba, but one giant leap for robotics.
Jones says his four-person company (plus consultants) was originally bootstrapped and is in the process of getting outside funding now. It might be a couple of years before the company is ready to deliver robots to commercial customers, however. “We’ve hammered down the biggest technological risks. There are no show stoppers,” he says. What remains is to make the robots more reliable, durable, and waterproof so they’ll work in the rain, heat, and other outdoor conditions.
At the MIT Enterprise Forum tonight, CEO Charles Grinnell showed video of a prototype robot in action. He also unveiled the new name for the company, which is in the process of being incorporated: Harvest Automation. Maybe it’s just me, but I prefer the intrigue of “Q Robotics.” Though I see that harvest automation could be a way to produce a bumper crop of new robot applications.
