The robots are coming
How long before there's a robot in every home? It's the stuff of science fiction right now, but Colin Angle, co-founder and CEO of iRobot, explains why that future is near.
IRobot, a privately held 12-year-old company, recently made a big splash when it introduced "Roomba," a $199.95 robot that can automatically clean floors for its owners. But company executives say far more exciting applications of robotics technology are on the horizon.
The 100-employee, Somerville, Mass.-based company intends to commercialize a host of other gizmos it's cooking up. The plan is to take products it currently sells to industry and to the military, and bring that technology to the mass market.
| I think you'll see multiple robots on the market next year for the home. You'll probably have three options to clean your floor at year-end 2003. By year-end 2004, twenty or thirty. |
Working with Johnson taught us a lot about how to clean floors and how to do it with a relatively expensive ($18,000) machine. We also worked with Hasbro to develop toys (My Real Baby). With those two partnerships, we had working knowledge of cleaning and mass-market products. By the time I got asked the question again, I said the time was right for (a product that) had to be at a price the mass market could afford. It required innovation on the cleaning technology and sensors to make it work at the right price point.
What's the story behind iRobot?
IRobot was founded to be a product company, and the challenge of making robots into products has been a long and circuitous journey. We have three divisions--one for the military, (one for the) industrial (sector) and one for the consumer. Creating tactical robots for the consumer has been a mission of the company's since we started the business. It's taken a while.
How did you start with the industrial products, CoWorker and MicroRig, and how many have you sold?
The CoWorker was a project where we started with a concept: If I could build robots with human-level intelligence, what would it do, and would it be a good product? The answer was yes, so we started to build one.
It started with the fact that we can't yet program a robot with artificial intelligence as good as a human's. So we asked, "What if I make a robot that's easy to 'possess' with another brain?" That lead to a standard Web interface (Internet Explorer) that would run the robot and allow the person to see and hear what the robot sees. The person would then lend his or her intelligence to the robot. The problem was cost. We developed with CoWorker what I call a physical avatar--the CoWorker is a robot that takes care of itself, but is controlled by a person sitting at the other end of a computer screen.
Where do you see it being put to use?
The applications can be very interesting--such as (using the robot to go on) tours of remote locations, where a person can lend expertise to the robot. If (the controller is) a doctor, the robot becomes a doctor and house calls are practical again. It's being used in industries where distance learning is critical, and having a physical presence and being able to point and look for yourself is key. It's used in the oil industry, where you can see half a world away.
What are the biggest industries for the CoWorker?
I can't say any one industry is more interested than another, but the oil industry is a good example. Any industry can benefit.
| Military robots will be a huge market within 10 to 15 years. |
The MicroRig is definitely in beta. It's a pickup truck for oil wells. It'll go into the well, carry payloads, help stimulate production and come back--without the need for the immense infrastructure typically required to put a payload down into the oil well. You don't need five miles of pipeline or tubing to push oil out to where the action is. It's in the process of being proven out...but it does take time, and the requirements are huge. The risk of screwing up an oil well has a high price tag. The industry moves rather slowly, just to be sure when they put (a device) in their wells that'll it'll do the things it's supposed to do. These products are the kinds of things where robots can make a real monetary difference.
What's the average price for industrial robots?
The prices aren't exactly set for these two robots. Because they are in the testing phase, they may need additional requirements that may need to be addressed. We're working with our partners to satisfy their needs, so the cost is shared right now on the development. That way it's something that economically works.
Is there anything that's out of beta other than the vacuum?
The PackBot is something we are starting to sell in the military. We're delivering units into the government labs and into the areas that will use the products. You can call up, and we'll sell you one.
How much?
It depends on how it's configured. It's between $50,000 and $100,000.
What are they designed to do?
At the base they have an extremely rugged platform that is controlled by a laptop over a wireless network. It has a computer in it. It also has internal sensors to make it drive where it needs to go to carry its payload. At the minimum, you put a head on it with eyes, sensors and lasers that would allow it to go in an area where you didn't want to go and send back intelligence.
How much abuse could it take?
It's a rugged platform that can climb stairs and take drops off (the) second story (of) buildings. It can take up to 500G shocks and is waterproof. Other than (from) Hollywood, there haven't been robots that have been thrown off buildings and kept going. In Afghanistan, it was driven off a 25-foot cliff into a mine and hard-packed dirt. You have a real rugged, fast device, which can be somewhere if you don't want to be. You can also have manipulation characteristics to pick up things to place payloads. There's a role for this in bomb disposal.
Can the PackBot withstand a bomb?
No. Not a serious bomb. You wouldn't send the robot to detonate the bomb. It may place something to detonate and then leave.
Where do you see robotics heading, and what has to happen to get there?
The future of robotics is the development of increasingly sophisticated products that deliver the same type of price performance and abilities. The PC started off as a curiosity, but once it became useful and had more productivity than its price tag, it took off and the rate of innovation took off. For the PC, the rate of innovation was slow until that economic hurdle was passed.
The computer was doing things no one has ever done before with financial tools such as spreadsheets. With Roomba and robotics, you're replacing some kind of physical labor. The kid next door is cheap and does a pretty good job, so the challenge of a robotic appliance is one that demands a true understanding of the job at a price point that's comparable to having the kid next door do it.
Is the mass market the holy grail of robotics?
The mass market is where the money is. Military robots will be a huge market within 10 to 15 years. It will continue to grow, but it's going to be a slower development. The PackBot is the first (robot) ever used in a combat mission. There's the start, but that's going to be more evolutionary than revolutionary. The consumer market will have rapid change.
How rapid?
I think you'll see multiple robots on the market next year for the home. You'll probably have three options to clean your floor at year-end 2003. By year-end 2004, twenty or thirty. The first shot is fired, and now everyone is scrambling.
Sounds like a glut of robotic vacuum cleaners. Are there other chores?
The mission of our consumer group is to make housework something that is done by choice. Our mission is to look at the tasks today and see what is robotizable and then work on developing the performance required to do it--whether it's washing your car, mowing the lawn or cleaning the bathroom. It's a huge and exciting engineering process, making the house take care of itself.
I imagine my bathroom will be more complicated.
It's a different process because there's water. I would say cleaning a toilet is different than tile. To do both is a complex challenge. We have research studying geckos and wall climbing. The big question is: Do we have to clean floor and walls, and will people pay for both? That's a matter for focus groups. Right now, it's difficult to say how it'll be.
