Few people have done more to bring humans and robots together than Rodney Brooks. Two decades ago, the Australian inventor, mathematician and former MIT professor founded iRobot, the company that designed Roomba, a line of robots that zip around homes and clean dirty floors. Today, he’s still dreaming up clever ways to make robots do our dirty work — but in factories rather than living rooms.
In 2008, he founded the Boston-based Rethink Robotics, a company building collaborative robots like Baxter and the one-armed Sawyer. These “cobots” are working next to humans in assembly plants and warehouses, handling many repetitive, dirty and difficult tasks. Brooks serves as chairman and chief technology officer of Rethink Robotics, whose investors include Bezos Expeditions, Goldman Sachs, as well as GE Ventures.
We recently caught up with him at GE Global Research in Niskayuna, New York, where he opened its Robotics Symposium. Here’s an edited version of our conversation.
GE Reports: How do you convince factory workers to start collaborating with your robots?
Rodney Brooks: Most of our customers are putting robots in places they never had robots before. Traditional industrial robots require a cage around them so people can’t get close to them because they’re just not safe. Our robots are safe. People can place robots into workspaces right next to humans and have them take over the really dull, repetitive parts of the jobs that people don’t like doing.
GER: How do you design them so that they don’t knock someone over?
RB: Our robots have force sensors in every joint. As they’re moving, they’re predicting how much force they should feel, and then, if they hit something, within a millisecond or two they’re aware that the forces are not what they expected. We quickly shut down the motion and then we go into what is called squish mode, where a person can just push the arm out of the way.
GER: Squish mode sounds awesome.
RB: Indeed. As a result, you’re never going to get trapped by the robot and you only get hit very, very gently. For my robots, I’m always willing to put my head right in front of the robot and have a whack.
GER: What else is driving demand for robots?
RB: In the U.S., in Europe, in Japan and in China, we’re already seeing manufacturing suffering from an aging work force. We just can’t get enough young people to come and do dull, repetitive jobs, in any of those countries.
We are aiming at alleviating some of that labor shortage by taking over the duller parts of the jobs. But it’s not just factory automation, or factories. There’s a labor shortage coming. There’s an incredible wave of shortage in farming in the U.S., for example.
GER: How are you developing new applications for your robots?
RB: At Rethink Robotics, we’re concentrating on factory and fulfillment right now, but we’ve also made the robots available with special software as a research tool to universities. We have them in over 400 research labs around the world. We hope that lots of smart people will use them and help us go into hospitals, into elder-care facilities and even food production. We don’t think we can solve all of those things right at once, but we can provide a platform, which lets researchers figure a lot of new stuff out.
GER: When we get old, we could have a robot companion?
RB: I think the elderly want to maintain their independence and dignity and age in home as much as they can. So as people get more frail, what are the robotics systems that can help them maintain their independence? For instance, a lot of people have to go to managed care when they can no longer get into or out of bed by themselves. A robotic solution seems like a good idea. But now you have to have a robot manipulating a human body, which is a very fragile thing.
Our robot arms are able to measure forces and they could be a good basis for a system like that. The ultimate solution is going to look different from our current robots, but it will be based on that technology. I think that I as I get older, I will have robots hugging me quite a lot.
GER: Have you ever hugged a robot?
RB: Of course, haven’t you?
GER: Besides hugging, what skills are you developing for the next generation of your robots?
RB: From the beginning, as we’ve been developing our robots, we’ve been trying to identify the friction points that stop people from deploying robots in factories. Early on, it was safety and having to put cages around them. Later, it was having to write lines of code for them. We brought those capabilities into our robots and you can now program them just by moving their arms.
Right now, our robots are doing a lot of packing into boxes. The next thing we are doing is giving them some elementary use of force connected to vision. It’s very easy to show the robot visual tasks, say, to look at a conveyor belt and pick up something from it. But we want our robots to be able to use force feedback as they are inserting, say, a delicate component into a circuit board, and also make it easy for the humans to show the robot what the exact parameters of the task are.
GER: What about awareness?
RB: It depends on your definition. When humans are working in a factory or any work site at all, they notice when something is wrong, and they raise an alert. You know, water pouring out of a pipe, a table is falling down, they know the fire’s somewhere. One thing we want to do is make our robots understand what’s usual and what’s unusual in the environment so they can at least alert humans.
GER: How big is the business opportunity for robotics?
RB: The Boston Consulting Group says that only 10 percent of automatable tasks in factories have been automated in the 50 years we’ve had industrial robots. The reason we haven’t seen more is because up till now, when you put a robot in a space, you have to get rid of the humans. But with collaborative robots we can mix humans and robots together. That helps open up somewhere between the current 10 percent and 100 percent of automatable tasks. There’s an incredible range of new possibilities coming about.