BotJunkie Interview: Nancy Dussault Smith On iRobot’s Roomba

Writing by Evan Ackerman on Monday, 17 of May , 2010 at 4:43 am

The home robotics market has gotten a bit more diverse in the last year, and iRobot’s Roomba is facing some new competition. BotJunkie talked with Nancy Dussault Smith, Vice President of Marketing Communications at iRobot, about what’s keeping the Roomba the robot vacuum to beat. She discusses how iRobot has been improving the latest generation of Roombas, how the cleaning behavior of the Roomba is not random, and how she sees the Roomba as distinguishing itself from other vacuum robots such as the Neato XV-11.

The entire interview follows, after the jump.

How did iRobot get started with the Roomba?

A lot of it was about the application, finding what we thought the right application was to go out with a practical robot for the world. We had a couple engineers who had gone to Colin [Colin Angle, CEO of iRobot] and said ‘we have this idea, we think it would be great.’ We did a little bit of research on it and went ahead and built some prototypes and tested them and all of the reaction we got was very positive. I think we initially underestimated the excitement around it and were pleasantly surprised to see how quickly it took off. We were optimistic, but I think it caught us by surprise how exciting it was.

Over the past generations of Roombas, what has iRobot been working to improve the most?

The very first generation of robots were made to vacuum, and when they were finished, they’d just shut themselves off, and one of the biggest problems was that people wouldn’t know where the robot was… They might turn it on before they left and then come home and not be able to find their robot for three hours. We learned a lot from that, once we started putting them in people’s homes, and found that a home base would be really helpful for the robot to be able to find its way home to the same place to be able to recharge itself. So, that was a big improvement on human robot interaction, I would say. It’s the same thing with the buttons on it: we use to have you choose ‘small,’ ‘medium,’ or ‘large’ based off of the size of your room, and now it’s simply ‘clean.’ We’ve optimized the robot’s algorithm to ensure that it can figure out and adapt to the environment that it’s in, so that it cleans for the appropriate length of time and gets everywhere.

With the actual cleaning, we’re constantly looking at ways to make the robot clean even better. The new ones have the dirt detect: [the 500 series] has a sensor underneath the brushes and it can detect dirty areas and stay there longer. Typically the area in front of your couch, or by the front door, or where your dog or cat sleeps tends to be a little dirtier than an area of your room that nobody ever goes to, so being able to spend more time at those areas is very helpful.

Looking at how people actually use Roomba in their homes is amazing. We’ve got over 5 million of these robots sold, and we have a great community of users who like to share feedback, their excitement about it, what they love about it, and what they wish was different about it. It gives us great little hints about how it’s being used and has taught us how to make improvements to each generation of product.

With that in mind, what do you see as the biggest potential area of improvement for the Roomba?

It’s a technology that I think is constantly evolving… I can’t speak to products that we’re going to be coming out with, or variations, but we’re constantly learning and evolving, and I think that areas of improvement for people consistently seem to be around the times that you use it. In the US, there’s a lot of really large homes, and people want it to run and clean the whole day, and so battery life is always an issue when you have something that’s powerful and cleaning and has to be recharged, so there’s a lot of time and energy that we spend optimizing its length of time cleaning and how many rooms it can cover intelligently.

Of course I have to ask about what we might be able to expect from the next generation of Roomba, or what other cool stuff you guys might be working on…

I wish I could talk about it, but I can’t, particularly as the competitive landscape heats up… We’re committed to making robots that work for people in their homes and make a difference in their lives, so there’s a lot of really exciting and cool things going on here, we just can’t talk about them publicly.

[I tried! -ed.]

A big part of that competitive landscape is the Neato XV-11; what do you think of their take on robotic vacuums?

I think when we look at the landscape as a whole, we launched in September of 2002, and people were thinking that within a year, or two max, there would be good competitors out on the marketplace, and it didn’t happen. Robots are really hard. If they weren’t hard, there would be more players in the game right now. They’re very difficult to figure out. And the hardest thing about robots is getting robots that can work in all different environments. So, when you look at your living room versus my living room, they’re probably very different: different size, different layouts, and they’re dynamic, they’re changing as we live in them. One day, when you sit down in your chair particularly hard after a rough day of work and it moves back an inch, your room now looks different to a robot that may have mapped it, and that can confuse it. So, when you look at how robots can adapt and work in those real world environments, that’s a really hard challenge, and so we have a great technology that allows us to do that, and we have a lot of great patents on those technologies, so it makes it difficult I think for some others to enter the marketplace because we have such great technology that we’ve been evolving. We’re excited about folks coming in with new technologies, but it’s important to understand the differences between technologies.

A lot of robots now are not using bump sensors, they’re actually trying to sense objects but not touch them at all. Say you had couch that had a skirt that came down to the floor. The robot would see that as a wall, and would not get underneath it, nor would it get behind it, whereas with our robot, it has a soft touch bumper, so it will slow down, and it’ll push it a little to see if there’s any give. If there’s give to it, it’s going to continue going. If there’s not give to it, it’ll back off. So now you’re able to get under the couch and clean there. If a robot is set up to be mapping a space, and it’s remembering the space, and that room gets moved even just a little bit, there’s going to be areas that don’t get cleaned because the robot thinks that there is an obstacle there when there isn’t, whereas with our robot, it has this ability to move around and actually figure out what’s in the space and how to clean around it.

Roombas are also set to clean in multiple passes. When we’re cleaning, we look at not just coverage, but also how well it’s cleaning while it’s covering. Those two pieces together I think make for the best robot cleaners, that can do a great job picking up the dirt, as well as getting everywhere, and we also know that multiple passes allow you to do a better job cleaning. If you think about when you use a standard upright, you go back and forth over things, and there’s a reason: it changes the nap of the carpet, it pulls things out… Going back and forth or going over things multiple times we know does a better job. Our robots are programmed to make sure that they cover an area three to five times before shutting off so it gives you a more in-depth cleaning.

There’s been a lot of talk about technologies where things are going in straight lines and looking like they’re really smart in that respect, versus our robot which is unfortunately… It breaks my heart when people say it looks random, because we have a bunch of engineers who worked their tails off on those algorithms. It’s not random. The robot is actually doing a lot of complex sensing and calculating and maneuvering, because in order to get into all those difficult places you need to adapt to the environment as it is, and our robot has the ability to do at. So, people sometimes jump to a conclusion as to how something is navigating, when in actuality, it’s something that we maybe just don’t understand because it’s operating differently than we do, and is just as (if not more) intelligent.

Anything new is exciting, and I’m not saying that straight lines aren’t sometimes good, sometimes, they’re great… If you have a wide open room with no furniture in it, that’s a great way to clean it, if there’s no obstacles. But not many people live in homes with no obstacles or objects in the room, so this idea that a robot needs to be able to adapt as it cleans is I think a really important one for people to understand.

Can you tell me more about the cleaning algorithm that the Roomba uses?

Our robot computes its algorithm 67 times every second, constantly stitching together information about its environment and recomputing its path. When it starts you’ll notice a spiral pattern, it’ll spiral out over a larger and larger area until it hits an object. When it finds an object, it will follow along the edge of that object for a period of time, and then it will start cris-crossing, trying to figure out the largest distance it can go without hitting another object, and that’s helping it figure out how large the space is, but if it goes for too long a period of time without hitting a wall, it’s going to start spiraling again, because it figures it’s in a wide open space, and it’s constantly calculating and figuring that out. It’s similar with the dirt sensors underneath, when one of those sensors gets tripped it changes its behaviors to cover that area. It will then go off in search of another dirty area in a straight path. The way that these different patterns pile on to each other as they go, we know that that is the most effective way to cover a room. The patterns that we chose and how the algorithm was originally developed was based off of behavior-based algorithms born out of MIT studying animals and how they go about searching areas for food. When you look at how ants and bees go out and they search areas, these kinds of coverage and figuring all of that out comes from that research. It’s not exact, obviously, I’m not saying we’re honeybees, but it’s that understanding of how to search out an area in nature that is the basis behind how our adaptive technology is developed.

Are there any other misunderstandings or things that people don’t really know about the Roomba?

The one thing I’ll throw out there is something about the cleaning. First of all, it surprises me that sometimes people ask me, “does it actually have a vacuum in it?” And yes, it is actually a vacuum. When people think about [vacuum] power, they think about it in the traditional sense of plugging in an upright, and knowing how many amps it has, and that means power. The way that the robot cleans is very different because it’s on the ground, and because it’s battery powered, we can’t have that type of power coming through the robot, it’s a different type of system, and we had to look at cleaning from a whole new perspective. The way that it works is with two counter-rotating brushes… They pick up the larger objects, so you don’t need that intense vacuum power to pick up, say, a marble. What you need the power for is the fine dust and debris. [We have] a very fine and thin opening for the vacuum port. It’s similar to a hose, when it’s wide open it just dribbles water, but you make that opening small and all of a sudden you have a tremendous amount of power. Same idea with our vacuum. [The opening] doesn’t have to be huge because all the big stuff is getting picked up elsewhere, and now you’re able to have great vacuum power with a lot less battery power, so it does a tremendous job cleaning.

The other thing I’ll throw out there is that we talk to folks a lot about this being a maintenance tool. We’re not advocating the mass burning of upright vacuums on your front lawn.. This is about maintaining a clean home, and having something that can do it for you every day, or every other day, not worrying about dust bunnies during the week. The [Roomba] will actually keep up with the mess. I like to think of it as a microwave to an oven: the microwave is to the oven as the Roomba is to the upright vacuum. It’s something you use more frequently, and you use it a lot, but it’s not going to cook your Thanksgiving dinner. It doesn’t have attachments that can clean curtains or get in between the crevices of your couch or do those types of things, but it does a really great job keeping your floors clean.

What excites you most about the future of home robotics and the direction of the industry?

The excitement is the potential. I feel like we are really at the beginning of an exciting industry. I get to go to a lot of different events and see a lot of upcoming ideas and what people are working on and the problems they’re trying to solve. I love the idea of manipulation, as well as robots working together towards common goals: the idea of a home that takes care of itself, were my robots all work together to maintain for me. If I can’t have Rosie, I want a swarm of robots maintaining my house, which I think is probably a little more practical.

What’s your favorite Roomba video?

The reason why [this clip] is my favorite is that I’ll never forget the night that it came out. We’d only been on the market a couple years and we were still small and growing. It was right around holiday time. I was at home, getting ready for a party I was having at my house the next day and I’m watching Saturday Night Live and that skit came on and I literally lept over my couch trying to get to my Tivo to record it. We’d hit pop culture! We were being spoofed on Saturday Night Live, it was such an exciting thing.

[ iRobot ]

Many thanks to iRobot and Nancy Dussault Smith for speaking to us.

Comments (4)

Category: Consumer

• Add this post to
• Del.icio.us -
• Digg

Name (required)

Mail (will not be published) (required)

Website

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>