By the time I was applying to graduate schools, my research statement was very much focused on this idea of wanting to become an astronaut and wanting to get a doctorate in something related to space robotics. It turned out that that year that I was applying, Rod Brooks at MIT had just started a planetary rover project, so it was a perfect fit between my interest and the new project. So when I visited MIT, I saw the autonomous robots in Rod Brooks' group; at the time a lot of them were modeled on insect intelligence. They would do things like follow walls and use simple image processing sensor techniques to find things like soda cans and pick them up and try to find a trash can. They were basically doing simple little behaviors. And when I saw that, I basically said to myself if we're ever going to see robots like R2D2, this lab is the place where that's going to start. This is the lab I want to be in because this is where not only do I get to do interplanetary rover stuff, but this is potentially where that vision of wacky Star Wars robots could really happen.

On making the switch to humanoids and socially-connected robots:
I think it was a natural progression where at the time they were doing planetary rovers, I did a lot of stuff that was biologically inspired. So in Rod Brooks' group, the thinking was to take nature seriously and to learn a lot from how natural systems are organized and so forth. A lot of my early work was with insects, but in many ways Rod's victory was to scale that up to more and more intelligent creatures. In 1993, he went on sabbatical and we knew that when he came back we would do something totally different, but we had no idea what. And he toured for a year, 20 different research labs all over the world, and he came back and said we're going to do humanoids. Of course we were shocked because we thought, you know, what happened to iguanas? What happened to dogs? Now we're going straight to humans?

One of the interesting things was that working with humanoids was really about bringing robots to the human environment — so much of our engineered world, our cities and furniture, it's really constructed for our morphology, the fact that we walk, we have two arms, and so forth. So if you wanted to build a robot morphology which is really well suited to a human-engineered environment, it makes sense to think about humanoids.

	Leonardo is Breazeal's most recent robot that interacts with human beings in a highly social, emotional way. CREDIT: "MIT Media Lab

The next big thing is how you would interact with people. Leonardo is Breazeal's most recent robot that interacts with human beings in a highly social, emotional way. And if you're talking about robots in human society, you're not talking about specialists, you're talking about the average layperson — Grandma, children, and people who know nothing about robotics, now interacting with these kinds of machines. That presented a whole set of challenges. I realized that the human environment is a profoundly social environment, and these robots are going to have to be able to do things not just independently of people but work with people, communicate with people, really be an integrated part of people's lives. Suddenly the emotional intelligence was very, very important because people are going to try to interact with these robots not as tools but as other animate life-like things. If you wanted to make the most natural interface possible, people are already pretty much experts at emotional interaction. The idea was to try to design robots that supported what we were already really good at rather than forcing people to learn a bizarre interface or a bizarre way of communicating to the robot.

Assistance and reminders from robots can help the elderly continue living independently. CREDIT: Carnegie Mellon University

The benefits of social interactions between people and robots:
There's an entirely different range of applications these kinds of robots could serve. So before it was like sweeping minefields and going to explore the ocean, now it's about doing things with people, helping people. In Japan one of the biggest applications or motivations for wanting to develop these emotionally interactive, intelligent robots is that they're concerned about their growing elderly population and the fact that soon they're not going to have enough young people to really tend to all the elder people. So they see robots as a positive technology to help the elderly live independently longer.

The elderly are often reticent about picking up a new technology, so it can't be something too confusing or esoteric. It probably has to be something that they see as genuinely helpful, but in the big picture people should actually really enjoy having these robots around as well. In many ways I think about a blind person's relationship with a seeing eye dog. The seeing eye dog performs a very critical function for that person, a very pragmatic, useful function. But on the other hand, people adore having their dog! So my vision was to use this social form of interaction to really address the needs of a person on a holistic level, not just about helping them with their cognitive and physical abilities, but also appreciate that people are social and emotional creatures and they have pleasure in interacting with things in this way.

- - - - - - - - - - - -
3 pages: | 1 | 2 | 3 |