Interview: Honda on the future for its Asimo robot
At the CES show this year, Honda’s humanoid Asimo robot was one of the big hits. He didn’t just walk and trot up and down stairs. He ran. He played football. He danced. And through it all, he cracked jokes like a Vegas pro. You can watch our videos of it here.
Naturally, Honda isn’t pumping millions of dollars into Asimo’s R&D to get a few cheap laughs. The company uses the bot in its TV ads for starters. But in the long-term, Honda wants Asimo to become a genuinely helpful companion for us humans. I talked to Stephen Keeney, North American Asimo project leader at Honda, to find out more.
CES seemed to go well. Was it a big deal for Honda?
Yes, it was a huge event for Asimo, as it was the first time we’d demonstrated it running outside of Japan. The reaction we received from everyone was so overwhelmingly positive too. First, they could not believe that a robot could actually run.
In fact, for many people Asimo is probably the first time they’ve seen a real humanoid robot outside of TV and the movies. And they start to realise that maybe someday this will be part of their life. People tend to say ‘Wow, I can’t wait until I can have one of my own!’
And he didn’t fall over…
That’s true! It doesn’t hurt that the people who attend CES are very technically inclined, so they wanted to see what the robot was capable of doing. Asimo is still very much a work in progress, and like all new technologies, once in a while we do have problems with it. But we’ve found it to be very reliable, and will get even better as time goes on, so we can get ready to introduce it to a home environment.
So what is Honda’s aim for Asimo in the long term?
Our goal is that one day it will be helping people maybe in their homes, although it could be working in a hospital or some kind of care facility for disabled or older people. So it’ll be able to get their medicine, answer the door, or provide them with a means to communicate through the internet.
That’s in the home. But maybe someday it will also do work that’s dangerous for humans, like fighting fires or working in unhealthy environments. The potential is only limited by our imagination, so if we can think of something we want it to do…
Is Asimo quite important for getting people interested in robotics too, or even just to persuade people that they shouldn’t be scared of robots?
Yes, definitely. We want to use Asimo today to try to inspire young people to become excited about the potential for science in their lives. We hope some of them will look at Asimo and say ‘wow, that’s so cool, I want to build robots like that in the future’.
The message is very simple: study any science you want: mechanical engineering, electrical engineering, software development, even chemistry to help build lightweight materials.
We also want people to understand that robots are not what they have seen on TV or in movies like The Terminator. These are machines that will really make our lives better in the future. We’ve conducted two national educational tours with Asimo so far, and it’s a wonderful opportunity to improve the image of science among young people. Too often, kids want to play their video games, but don’t want to understand the science behind how a computer is made, for example.
Are young people more receptive towards robots though? After all, they’ve grown up with them as part of popular culture, and not always in a ‘taking over the world’ sense.
Yes, younger people are very receptive to a robot being part of their lives. They’ve grown up in a culture of videogames and computers, so they don’t think it’s strange that a robot will help them do their homework or clean up their room. In fact, what they usually ask is whether he can go skateboarding with them! They have some amazing creative ideas about how Asimo can be a part of their lives.
Older people are not as comfortable, but they look at him from a utilitarian viewpoint, about it doing the vacuuming or going shopping. They can see the value of having something to do things for them and make their lives easier. We’ve designed Asimo to be very friendly looking, like a small child, so he’s not imposing in the home environment. But its big enough that it can do things, and it even has a smile on its face! People aren’t really scared of it.
So being humanoid will be important for these kinds of domestic helper robots? They couldn’t just be a box on legs?
It is important for the perception. Humans have a natural tendency to feel comfortable around other human beings, or things that look somewhat human. But also, it’s an important factor if we want Asimo to be truly capable of helping us.
Think about our world: it’s created to accomodate humans with two arms and two legs. Stairs, steps, doorknobs, lightswitches… Everything’s been designed to accomodate our physical being, so if we want a machine to work efficiently in our world and carry out the tasks that we would do, it needs to be created like a human body.
That’s why we didn’t work on a robot with wheels or treads. Of course, in the future, your home might have Asimo as a general helper, then other robots doing specialised work like vacuuming. But we view Asimo as a multi-dimensional robot capable of carrying out many different activities, rather than being specialised. So that’s why he’s more of a human shape.
What are the key technical improvements that lie ahead, now you’ve got Asimo running?
The big challenge now is to enhance the artificial intelligence. We need to make Asimo truly smart enough to understand what we want it to do, and then carry out that command on its own. It’s incredibly complex, getting Asimo to understand when we’re talking about a bottle of medicine or the TV remote control, where you’d find it and what it looks like.
We’ve perfected the mechanics of movement and mobility. Now we really need to work on Asimo’s smarts, so to speak. That’s what we’re focusing on now.
How important is connectivity – Asimo being hooked up to the internet and so on?
It certainly will be important in the future. Asimo can already connect to the internet, but someday we will need an incredibly large and complex database of information from which Asimo will pull information to carry out particular tasks. It would be most efficient if it could connect to some large server system or data bank to, say, look up medicines. Or if it’s fighting fires, it could retrieve information from a database of what other firefighters have seen. That kind of connectivity will be very important.
Since Honda started work on Asimo, there’s been an explosion in entertainment bots – the cheaper models that are a step up from remote-controlled toys, but are still made to play with, rather than as home helpers. How much crossover is there between the entbots and something like Asimo?
We are committed to creating a complete humanoid robot that’s really capable of helping us in a multitude of ways. Our first objective is a robot that can help people who can’t really help themselves, if they’re in a wheelchair or a sickbed. So it’ll move through their environment for them and carry out different tasks.
We’re not really focusing on entertainment bots or the lower end of the market. There are a lot of other companies focusing on that, and some are doing it quite well. I think it’s good for the field of robotics in general for all these companies to be involved. Roomba, Lego Mindstorm… all of these robots help make people more comfortable with the idea that robots will be a part of their lives one day.
This is an important step as we move towards the day when humanoid robots will be in our world helping us. But Honda is working on the higher end robots. There’s quite a big leap between entertainment robots and the ones that are really going to help us.
We want to take our engineering and technical expertise that we’ve developed through everything from motorcycles and jet airplanes through to Asimo, and build a very complex machine that will truly be a useful helper to us.
So, the big question. When can I have an Asimo in my house? When will these robots be on sale to the general public, and will they be prohibitively expensive?
It’s one of the questions I get all the time! A lot depends on the speed or progress of development in general technology, like processing chips becoming smaller and faster, and memory becoming more powerful and cheaper. So a lot depends on how quickly the technology can improve.
We think within about 10 years, we’ll see the first version of Asimo helping people in the home. It’ll be a gradual process though. Look at the first people who owned cellphones, which were pretty clunky at the time, but now everyone has one.
Asimo is a new technology that will continue to evolve, but we hope in about 10 years we’ll start to see versions working in homes or hospitals. I just hope by the time I retire, I’ll be able to have an Asimo!
As for cost, if you think about it, a lot of older people need help on a daily basis, so they have a healthcare provider coming in. If a robot could do some of those tasks, they would only need a person to come in once or twice a week. So even if the robot were the cost of an automobile, if you think about the savings you’d make over time, that might be a reasonable alternative, and health insurance might cover some of the costs. We think there will be a huge demand for it as the population ages around the world.
MORE ASIMO GOODNESS
Videos of Asimo at CES 2007
Our Top 10 suggested jobs for Asimo and his descendants