Drug-Resistant Fungus Computing's Top Prize Google's AI Chatbot Beat Airline Ticket Prices ChatGPT Bug 7 Daily Habits for Happiness Weigh Yourself Accurately 12 Healthy Spring Recipes
Want CNET to notify you of price drops and the latest stories?
No, thank you

Q&A: iRobot taps into its Warrior spirit

CEO Colin Angle talks about how robots are weathering the recession, what to expect from PackBot's big brother, and why iRobot isn't building a humanoid.

Corrections were made to this interview. See below for details.

The PackBot robot has made a name for itself in dangerous places like Iraq, but the future may belong to both its bigger and smaller siblings.

U.S. military forces have long made use of the PackBot to discover and disarm roadside bombs, keeping flesh-and-blood soldiers out of harm's way. Now its maker, iRobot, is looking to make inroads with two variations on the design.

iRobot CEO Colin Angle
iRobot CEO Colin Angle Jonathan Skillings/CNET

The SUGV (short for small unmanned ground vehicle) may be one of the technologies that emerges in good shape from the Army's massively expensive Future Combat Systems program, which seems likely to be significantly deflated as the Pentagon shapes its budget for the coming year. Lighter and thus more portable than the PackBot, it seems in good position to be spun out from FCS evaluation teams to other units.

Meanwhile, iRobot is beta-testing the much larger Warrior, which will be capable of carrying bigger payloads. As upbeat as folks at the company are about the PackBot and the SUGV, iRobot CEO Colin Angle describes the Warrior with unabashed enthusiasm: "That robot is going to change the world, and change the perception of what practical robots are all about. We're pretty passionate about that."

The Bedford, Mass., company has its more domestic side, of course. Where it's sold 2,000-plus PackBots worldwide, it has sold more than 3 million of its Roomba floor-cleaning robots. It's got a range of household bots, from the gutter-scouring Looj to the pool-scrubbing Verro.

Angle met with CNET News on Monday at the company's headquarters, ahead of this week's RoboBusiness conference in Boston, to talk about how iRobot--which had $307 million in sales in 2008--is weathering the recession, how robots are changing battlefield habits, and why iRobot isn't building a humanoid bot.

How is the recession affecting iRobot?
Angle: There's no doubt the recession is having an impact. We first started to see the impact of the recession in October of last year, so we were able to say, OK, the rules have changed on consumer spending, and we were able to adjust our guidance for 2009. So we hopefully have done the right thing, but domestic sales certainly are negatively impacted.

The good news is that internationally our sales remain very, very strong. And perhaps because we were operating off a relatively small base, relative to the potential opportunity, we continued some momentum. So the guidance that we were able to give year to year had domestic down and international up on the home side. And then on the military side we had a different set of non-economy-related factors slowing down growth temporarily.

With the international growth, how widely dispersed around the globe is that? How many countries are you selling into for both (home) and military?
Angle: On the home side, we are relatively broadly distributed throughout Europe, Korea, Japan, Australia, most major developed markets we're in at this point. On the military side, we have sold, we have presences in about (13) countries, so we continue to build on all fronts. It is a fairly significant number. The utility of these types of robots, in particular the PackBot, has been recognized globally.

There is always going to be, at least for a long time to come, a human in the loop, as far as trying to decide when a robot should employ lethal force. You know, AI is not to the point where a machine should be making a life-or-death decision, and I wouldn't even be able to tell you when I thought that might actually come to pass.

The PackBot seems to be a very flexible platform, there are a lot of different things that have been put onto it--IED detection, sniper detection--no weapons so far...
Angle: True. You know, it is a platform, and explicitly designed to be one, with common interfaces. We have the Aware 2.0 robot intelligence system, or software platform that has open APIs so that third parties can develop different payloads for the robot. We have a developer conference every year; last year over 80 companies attended to learn how to build and interface their products to our platform.

And the PackBot is, right now, the most numerically successful, plus the oldest platform. We have two new platforms that are starting to come online -- the SUGV, the small unmanned ground vehicle, which is a 30-, 35-pound platform. At 50-60 pounds, you have the PackBot, and then at 250 pounds, 300 pounds, you've got the Warrior, which is the larger size that is now going into beta testing and starting to get into the combat exercises to look at how it can effectively be deployed.

The vision is, iRobot will be the central point of portable ground robots, and depending on how big a thing you want to put on the robot, you'll work with us to deliver that capability, deliver it to the soldiers. The more these products are used, the more soldiers say, here's an opportunity to perform a mission more safely, to perform a mission more quickly. We'll see a lot more missions starting to be taken on.

The PackBot's not the only robot platform that size that the U.S. government uses. There's the Talon from Foster-Miller.
Angle: The Talon is heavier than the PackBot, so that's fine if you're getting to where you need to use it in a vehicle, but as soon as you start having to carry it, it's less so. Both the PackBot and the Talon are used extensively for this roadside bomb threat in Iraq.

I know from what I've read that the soldiers are very happy to have those devices.
Angle: You can imagine, because -- if you're in Iraq and there's a bomb over there and you know there's some way of detonating the bomb, you haven't quite figured it out and someone may be watching, and his entire goal is to explode it at a time when it can do the most damage to you, and your job is to defuse it, that's just a bad situation. This is giving a new lease on life to these guys. I've spoken to many of them and they all credit their own lives with our nation's ability to procure and build these types of systems. It's a godsend.

And now we're starting to see with the SUGV, and the robot's getting into the hands not just of the explosive ordinance disposal guys but the regular infantry, whose missions include things like, go clear that building. And many times of day as you're trying to clear building after building, you're faced with, OK, there might be something bad beyond this door. I can throw a grenade in and maybe kill someone, an innocent person, or I can jump in the door and get shot. Wouldn't it be nice to have the better option to have the robot go in first and evaluate what's going on, and improve the outcome?

The proposed DOD budget looks like it's favorable to things like the SUGV. From what I've seen, the Future Combat Systems--a lot of that probably will fall away, the bigger vehicles, the common chassis. But Secretary Gates and others seem to want to have more of the robots.
Angle: Well, this is something that's far from played out, but certainly from what has been said, the SpinOut 1 technologies, in particular the robots, have a place and could come out of the (budget) actions, in fact, enhanced, and find more resources associated with them. Anytime there's change, there's risk. But from what's being said, and the type of combat the United States finds itself in these days, the robots are on the rise as far as being an important part of the new enabled technology to keep our soldiers with that advantage.

Some people, including (author) P.W. Singer, have raised the issue of battlefield ethics, the laws of war, and how robots might change that. How do you see the use of robots reshaping combat?
Angle: There is always going to be, at least for a long time to come, a human in the loop, as far as trying to decide when a robot should employ lethal force. You know, AI is not to the point where a machine should be making a life-or-death decision, and I wouldn't even be able to tell you when I thought that might actually come to pass.

But what a robot does give a soldier is the ability to shoot second--which is, again, incredibly empowering and important. Let the robot be the thing on the point, let someone attack it, and by doing so reveal his position and reveal the intent of what they're doing there. This is a big, big concept--robots let the soldiers shoot second. Additionally, a robot can carry nonlethal technologies, where a soldier is much, much less willing to do so. If a bad guy's over there with an AK-47, I want my M-16, I want my machine gun.

(Robot technology) is something that you need to be very careful of. It also offers some of the best hope for humanely dealing with this new type of bad guy that we're faced with.

And the old types, too.
Angle: The old types, too. But we're not going to have tank battles in the fields of Eastern Europe where everybody lines up and you know what side everyone's on. That's just not going to happen anymore. So we find ourselves in a situation where it's much more difficult just to understand what you're supposed to do, much less do it, and that requires either a lot more soldiers, in which case we're putting a lot more American lives at risk, or some kind of new technology can go in and serve the function that those soldiers served. And that's where, I believe, robots come in.

It seems to me, too, that popular perception gets ahead of itself in thinking about robots which, in movies, are very autonomous. Most of the robots that are used in battlefields--the PackBot, the Predator--they're still mostly, if not entirely, remote-controlled. Somebody is still driving that machine, and it's not working on its own.
Angle: There's a man in the loop, and even if you could tell the robot with the GPS, go that way, there would still need to be a person in the loop to decide what to do once the robot got there. There's going to be more capabilities built into the robot, so that a soldier doesn't have to his head down looking at a video screen all the time while someone sneaks around and can mess with him, and in that way we'll allow robots to be more effective. But there needs to be a person in the loop, because AI is just not that good.

People don't understand that before we have a robot human, people are going to be doing things like incorporating robot technology into their own bodies.

Hollywood loves to portray the Terminator, or loves to say, here's the future, and have the little boy knock on the door and "Hello, I'm a robot." It's just not going to happen. Maybe one day, but by the time that happens, the world is going to be a far weirder place. But people don't understand that before we have a robot human, people are going to be doing things like incorporating robot technology into their own bodies. And the ethical challenge for the field of robotics is going to be answering questions around what happens when you can have a neural interface to the future of the Wikipedia, and so if you have this capability (built into your body), which probably would cost money to do, you'd have a huge advantage over someone who didn't--now there's an ethical question.

Right. It's sort of, cheat notes writ large.
Angle: Right. Everyone unplug their neural implant before they take the test. Right now there's medical procedures to help the deaf with cochlear implants, and there's some artificial eyes that are starting to allow blind people to see some amount of image, differences between light and dark, and it's getting better all the time. But what happens when that becomes elective surgery? Do you want to have an iPod embedded in your ear? Do you want to have your computer screen--you know, remove a good eye and put in something else? I mean, these questions are going to be fascinating and causing us to really dig deep as to what is right and wrong, far before we have to worry about true, artificially intelligent robots beating us on the battlefield or in society.

And both those (scenarios) are a long way off--robots powered by very smart AI, and humans having reached some sort of cyborg threshold.
Angle: Well, I think that the second is far sooner than the first. We are already doing surgery where we're implanting machinery into our bodies. Dean Kamen has developed a fantastically successful prosthesis, with a neural interface. They've figured how to attach it to your nerve bundles so you can think and the robot's arm can move with very little training. I mean, you talk about what's going to happen sooner, well, we're already seeing some of these cyborg-esque futures becoming more and more real. I think it's going to be a fascinating place, but I think that the typical concerns about what happens when you weaponize robots and what happens when we have robot people hanging out amongst us are kind of the easy but wrong questions.

On the home front, the Roomba is still the center of focus for iRobot?
Angle: The Roomba is certainly the largest revenue driver at this point. It has a head start, it was the first thing that we did, and while the brand iRobot entrusts that the products we make are truly practical--they're not gimmicks and actually work-- that helps the subsequent products come up quicker. The Roomba still has a relatively tiny penetration--we've sold over 3 million of these robots. It sounds like a big number, and it is, but compared to the number of households in America, it's a tiny number.

Our users are very passionate about the product, they tell their friends and so forth. We'll continue to see, just driven by our installed base, more people, more success, driving more Roomba sales, and then with sales of the other products moving along nicely, but still in the shadow.

We keep at it. We're not done. The mission of the division is to keep working at robots that will help tackle the dull, the dirty, and the dangerous--the routine maintenance tasks that we're faced with, and once we're done with that, trying to turn our focus to the people who live in the homes and with this notion of helping people live more easily, more independently. It's early days. You know, we can vacuum. We can vacuum well, and scrubbing is coming along and so forth. But there's so much more a robot could do as far as helping you come home to a house that is exactly the way you want it, with no need for you to go and do these maintenance types of tasks.

What would be the next thing to tackle?
Angle: Well, we don't actually talk about what we're doing next. It's sort of, in the future here's the body of things.

Still no lawn mower.
Angle: There's the lawn mower, there's cooking, there's windows, there's more stuff going on in the bathroom with your tub, and doing laundry, folding laundry, putting stuff away. Once we get manipulation on the robots at consumer price points, those are all very real, very doable sorts of things. Shoveling the driveway.

What about areas like robotic surgery, robots for NASA, something farther afield like that?
Angle: I think that robots for surgery is a great application for robotics. The Da Vinci system from Intuitive Surgical has really shown that robots can augment regular doctors, not just for tele-surgery, but in the same room, give that doctor more arms and do more precise movements. This is fantastic. It's not an iRobot area of expertise but it's certainly a real driver for the industry.

Industrial cleaning is going to be an industry robots take by storm once the right product comes along. The exploration of the solar system has already largely been given over to robots to take on--the Spirit and Opportunity on Mars, the probes we keep sending out. These are demonstrating that if--ultimately you're going to want to send people because it's more fun and interesting to go start a colony on Mars. But the prep work in making it viable and feasible to know what to do, that's the work of robots.

Why would you want to make a humanoid robot? I mean, I guess for making movies they're good. If you want to have a robot companion, maybe it should be humanoid. But other than that, most tasks are best tackled by designs that are not constrained by trying to look like a person.

What about a humanoid robot?
Angle: Why would you want to make a humanoid robot? I mean, I guess for making movies they're good. If you want to have a robot companion, maybe it should be humanoid. But other than that, most tasks are best tackled by designs that are not constrained by trying to look like a person. I mean, balance and walking are incredibly hard things to do. If you look at some of the Japanese walking robots, because they're very focused on solving this problem, and then compare it to Warrior, our large, dual-track system, and say, OK, which one makes more sense?

The Asimo (from Honda) requires a team of 10 or 15 people to maintain it, it can walk about, maybe, half a meter per second and in some situations climb stairs over the course of a few minutes, and if it ever falls down, it's a paperweight. I think it has something like 40 or 50 motors in order to make it work. Then take the Warrior, and the Warrior can take a 10-foot drop onto concrete, drive (10 miles an hour), drive up stairs without stopping at full speed, carry (150 pounds) of payload and has, maybe, five motors. So it's stronger, it's faster, it's more durable, it has more efficiency, and it can go nearly everywhere a human can. So you look at these things and say, which one of these is a robot human? The answer is, Warrior. Warrior is designed to operate in environments that we have designed for ourselves, as efficiently and capably...

Roadways, buildings...
Angle: Right, so if a person can do it, chances are Warrior can do it. And we're developing arms and all sorts of payloads to allow it to take on some of the most dangerous tasks that we currently are forced to give to ourselves to perform. That robot is going to change the world, and change the perception of what practical robots are all about. We're pretty passionate about that.

Obviously there's a lot of in-house R&D that you guys do. But where do you look for new ideas--there are a lot of robotics labs, Carnegie Mellon, MIT, other places.
Angle: You're answering your own question. I mean, we have a technology road map as to where we want to bring the technology, but we also take very seriously our academic ties--Stanford, Carnegie Mellon, MIT are three of the leading places that we make sure we know what's going on, because they're a fountain of good ideas. We go visit campus, and hire students. It's a community that likes to get together, talk about what they're up to. Building robots is extremely difficult and trying to just go off and do it alone is (daunting). You have to be humble in this industry.

Helen Greiner and Rod Brooks both left the company last year, they were two of the three co-founders. How are things different nowadays without them here day to day. I know they're on the board...
Angle: They're on the board, and that's the key thing. Rod, when he was at the company, was playing a role that is helping to network, helping to keep us aware of what's going on in the industry, and that's something he can do, maybe a little less formally, but certainly very effectively, from a position on the board. Certainly Helen leaving the company--she's a very talented individual and we miss her, but she's still intimately tied into the business. What she's doing is synergistic with where we're going with our military division.

I wish them well and success. It's difficult to predict exactly what's going to happen, but it's a boon to the industry. And iRobot still gets to tap into the association with these two pioneers. So it's a little bit different, but it's not like they're gone, and we've been able to continue to grow and thrive and things are still very, very exciting. We've been able to attract a lot of new, brilliant roboticists to be at the company where we are making it happen and putting actual robots in the hands of soldiers and customers around the world.

Correction, April 16, 11:19 a.m. PT: This interview initially misstated the number of countries into which the PackBot has been sold, along with the speed and payload for the Warrior.