|
||
|
||
|
|
|
|
|
|
|||
|
|
|
|
|
|||
|
|
|
|
In this Gartner Fellows interview, conducted in October 2009 we explore the views of a chief executive and thought leader who is in the driving seat of the rapidly growing mobile service robots industry. Interview conducted by Mark Raskino
Mark Raskino (Gartner): What does iRobot's business story tell us about the robots industry? Would it still be a 15 year journey for a new entrant today? Colin Angle: It would be much easier, but it's a difficult industry. The designing and developing of robots is analogous to designing a car, or some other highly complex electromechanical system. There's a lot of IP, you have moving parts, you have [product] liability and quality. It can be very challenging. Your margins are oftentimes difficult to maximize – you're not a software business. You have to be expert at manufacturing. But on the other hand the nice thing about the industry is that it is still a nascent industry, I once had a guy come up to me in a conference and say, "Well gee Colin, y'know you've done vacuuming robots and bomb disposal robots. You've done it all. What have you left for the rest of us?" And I kind of laughed and said – "You have no imagination." So as we look at really what is going to drive this industry, there's a lot more to play out on the military side: intelligent vehicles, the auto industry is huge, having robots do more deep oil exploration and help us solve some of our energy challenges is a fantastic market. The industrial need to maintain and clean buildings is a multi billion business. And at least in the States it's something that's constrained by the availability of labor, people willing to do it. It's a fantastic industry. The medical industry is a huge thing. We're at this interesting point where the baby boomers are starting to get old. The nursing homes and assisted living facility industries are in decline, we're racing for a brick wall as far as how we're going to care for these people. And then there's just finding ways of making our homes more able to maintain themselves, so we can focus on other things in our lives. Gartner: In 10 or 20 years time, will we see the equivalent of a Toyota or the IBM arise in the robot world? Do you think there will be corporations of that scale? Angle: I think in that timeframe, absolutely and it's up to the will of the emerging robot business to decide which growth path they want to take. If you go back into the PC industry there were companies that had they been acquired would have led the landscape to look very different today. The question is, at what price, and when do you do it. Do people really have the vision to say, "I would like to do that?" Would the emergent company thrive and succeed inside the auspices of a larger company. Because oftentimes, especially in businesses where a lot of investment is required and a lot of risk is required in order to invent the next great thing – large businesses have a less successful track record than a more aggressive, less risk-averse, smaller company. And then once a smaller company finds that, they instantly become a more expensive and thus more intimidating entity. The 'Wild West' breadth first search of opportunity during the period of maximum entrepreneurship has now been focused down into making the greatest robots and the most appropriate robots for our customers that we can, and get that economic engine really started and humming along in a great and sustainable fashion. One of the reasons why we went public was to have a currency for acquisition and to grow inorganically as well as organically, because we felt like our success would make it easier for people to come and follow and that we could take part in a growth model that is proven that in many other industries where there's going to be a lot of fast followers who come up with interesting technology, and then if we find nice adjacency, or we find nascent products that need channel and manufacturing, [and using] the sophisticated processes that we have, we could take those companies on and then see a nice return for investors based on expanding our product line. So for example we've made an acquisition last year where we're adding unmanned underwater vehicles to our Government and Industrial Robot division's product line. Gartner: So it's a story of counterbalancing and rebalancing as you evolve a new industry. Angle: Right. And it keeps going. The government has invested approximately $20 million a year [with us] in sponsored research and development. And so in order to stay ahead of the technology curve, competing with other large consumer electronic companies, we were able to bring in sponsored research which when you add it to our internal ability to put dollars against new products, you're in the mid-teens as far as effective R&D research, which given our lower margins because we're a product company, not a software company, allowed us to support one of the largest robot research groups in the world. And that helped our position and helped our technology. So over time, technology developed on the government side goes and help the consumer side, and vice versa. For example, the track design on our gutter cleaning robot was informed by track work on military models. Gartner: Does any learning go in the opposite direction? Angle: Yes, we have a development program called 'Landroids'. Soldiers want a really very tiny, inexpensive, disposable robot that they can just throw through a window and our experience of doing low-cost manufacture [for consumers] has enabled us to bid and win. We will incorporate some of the technology into this lower cost 'throwable' robot business. Gartner: These small throwable objects, what would their function be? Angle: The throwable robot is more in the ideation, starting to see some prototypes and demonstration stage. So I want to put that in context, but the idea is that every soldier would have one of these robots in a little pouch on their leg or in their backpack and what have you. And if they wanted to leave something behind so no-one snuck up behind them, or they're at the top of some stairs and want to toss something down the stairs to see what's down there. It would give them a personal aid. The original impetus was to toss them into a building and they would spread out on their own to create an ad hoc [communications] network inside a building – but once you have the robots like that, then everyone's coming up with new ideas about what they can do. Gartner: Bill Gates wrote an article in Scientific American in 2007 suggesting that standardization of robotics software development platforms might create explosive growth as he saw and helped to create in the PC industry in the 1980s. Do you broadly agree with that? Angle: Well we talked a little earlier about the challenge of entering the robot industry. There's just so many things you need to get right. The way to make it easier is to have sufficient penetration of a platform that has open interfaces. It would be very hard for the software industry to succeed if they couldn't develop apps that ran on Windows, that ran on PC, and generate revenue. And so that model basically holds but the challenge for robots is that the integration of software and hardware needs to be much more intimate because of the mobility – there is a much tighter coupling. I think that over time there would be space for a pure software play. But that's not today. What the new industry needs to succeed today is an installed base of an extensible platform with open APIs. Gartner: Vacuum cleaner engineer and entrepreneur James Dyson said in a New Scientist interview this year: "We made a robot back five years ago but we didn't launch it because with 85 sensors, powerful motors and huge numbers of batteries it was just too expensive." Angle: Dyson, he's a brilliant man, he's made some fantastic vacuum cleaners and he's a great marketer. But his vacuums are simple compared to what a robot vacuum needs to be frankly. Worrying about navigation and doing systematic cornrows or whatever on the floor doesn't necessarily speed up how fast you vacuum, and creates huge problems of systematic neglect. We have sold over 4 million home robots and every room in every house is completely different. So a systematic strategy is going to fail. What it sounds like he did was put a lot of energy into some very sophisticated expensive sensors that try to over-control what the robot was doing, and that cascades into a big fat expensive robot. Gartner: So you want the simplest path to solve the problem. Angle: Yes – probably 85% of the software on Roomba is about how not to get stuck in a corner, rather than using intelligence to recognize a repeating pattern from expensive sensors. These are lessons learned out of being in an industry for 20 years. What a customer wants is to take the robot out of the box, hit 'Clean', go away, come back, and find the Roomba has plugged itself back in, is recharged, and the floor is clean. Anything that distracts from that mission adds cost. The other problem with adding stuff to your robot is they get bigger. And the most expensive part of the robot is the battery. If you double the dimension, you increase the power by four times and all of a sudden again you're into $1000 robot. Gartner: The Roombas are autonomous but your military devices are mostly remote controlled by human beings. So what is a robot? How can we define and segment that term? Angle: Mobile robots are a class of devices that perceive their world, 'think' about what they see, and take physical action based on what they think they should do. That's a simple definition. The expertise of a robot company is about the integration of information from sensors using artificial intelligence capabilities to get the physical platform, to actually do what you set out to do. And that's a huge problem – we worked with Johnson Wax for nearly four years to become experts in cleaning. Domain expertise for what are you trying to do, needs to be part of the equation as well as the more technological aspects. So to be a generalized robot company you're going to need to somehow acquire a lot of different domain expertise. Gartner: This year, British academic Noel Sharkey has called for international public policy debate about ethics in the context of offensive military robots and the kill decision. Angle: It's a complicated question and it's one that is discussed in theory because we are ahead of ourselves. Let's be honest. The notion of having three laws of robotics is a fanciful notion. It was a plot artifact. It's going to be more than 20 years before a robot has the ability to understand enough about the environment that it's in, in order to have ethics, in order to make moral decision about whether or not to pull a trigger. This is something that is incredibly complicated and subtle I don't begrudge people thinking about it, talking about it, but think the future is going to be far stranger than Terminator robots. I think the real hard moral ethics are when people start incorporating robot technology into their own bodies. If your child decides to have elective surgery to remove their good eye and replace it with a robot eye because it has monitors and it has x-ray vision and has cameras built in and so forth and it looks identical to your eye – that's an area where I would encourage people to spend some mental time around. Robots are not going to be moral devices for decades. And so that the one thing that we need to think about when weaponizing robots, is that a man is always in the loop. The U.S. has a very strict doctrine around the fact that if you want to use a robot, a human has to be involved in order to activate and make a lethal or near-lethal decision. So that's important. Gartner: Thank you for that very thoughtful answer. If you will forgive me I'd like to end this very interesting interview with a couple of trivia questions. Angle: Go ahead. Gartner: Did you ever meet Isaac Asimov? Angle: No I did not. Gartner: That's sad. Have you met Arnold Schwarzenegger? Angle: No, but our French distributors did. In fact one of our guys in France has an Arnold motorbike from the original movie set. Gartner: What about Will Smith? Angle: We like Will. We worked on iRobot the movie with them, and so actually went to the premiere and hung out with Will for a while. But I'll tell you I have a problem with iRobot the movie. They've got three main characters – who are the equivalents of our three founders here at iRobot Corporation – Rod Brooks our scientist, Helen Greiner and myself. In the movie the scientist dies, the CEO dies and Helen ends up with Will Smith. So y'know, I'd like a movie with a better outcome for the CEO of the company.  |
|
|
||||||||||||||
|
||||||||||||||||||||
|
