Sebastien is a French student studying in Japan for his masters in Robotics.
Sebastien, tell us a little about yourself and why you decided to work with robots.
Currently, I am a master student in robotics at the University of Tokyo, a world leader in robotics research and technology.
I was drawn into robotics by my father mainly. He built his own computer on the living room table in the early 80s, and I loved to make RC models with him. Later when I saw the first images of Spirit and Opportunity on Mars, I just wanted to do that.
Since I also wanted to travel, see the world, I went abroad for studies as soon as I saw the opportunity. That’s why I am a French student in Japan now.
How hard is it for a non-technical person to build a hobby robot? What should they expect?
It is very confusing to start into hobby robotics if you don’t have anyone to mentor you. This is because robotics is the convergence of 3 very different fields: computer science, electronics, and mechanical design.
If you want to start softly in robotics, you should learn them 1 by 1 if possible. I think kits make them easy. You can learn electronics with Arduino, programmation with Lego Mindstorms, and mechanical design by just building new shapes every time.
I also have a website, http://www.about-robots.com, that I want to use to help people getting started in robotics. It’s not that hard anymore, if you can find the right information.
Concerning humanoid robots, what challenges are engineers having to progress to the next level? What are some of their thoughts about overcoming these challenges?
Humanoid robots are still barely walking yet. There is a lot of research needed to make robots walk on 2 legs. It’s about everything: they design new hardware like hydraulic pumps to replace electric motors; they need new algorithms to keep balance, and better Artificial Intelligence to make them more autonomous.
But there is another fundamental issue with humanoid robots: who needs them? Do you really have enough places in your home to accommodate another member?
If you have read the “Robots” series by Isaac Asimov, you see that humanoid robots were designed to work in space station, or on planets where the density of population was extremely low. There are no robots in over-populated cities…
The few successful humanoid robots existing are designed for very specific tasks, and none are made for large scale production yet.
In your opinion, how far are we away from creating a robot that can hold conversations on its own? Please explain.
Actually, not far. This is about Artificial Intelligence more than robotics, although the application is obvious.
Every year, chatbots (computer program designed to chat via computer with you) are competing to pass the Turing Test. Many chatbots are able to make a sensible conversation for a few minutes, although they are not “believable” yet. But the progress is good.
On the other side, if you want to talk to a robot, you need the robot to have “speech recognition”. This is what Siri from Apple uses to understand your commands. This is also what Google Speech-To-Text technology is doing.
Nao, a small humanoid robot designed for robot interaction research, uses the same speech recognition software than Siri. Several programmers are also using Google’s tools to make Nao understand your sentences via the Internet.
So talking with robots is really coming soon, I guess.
What is currently the most advanced robot in the world? Tell us about it.
There are so many different robots; it’s hard to tell which one is really the best one. So let’s pick 2 of them and see what they have that put them on top.
The Google Car, and the many equivalent vehicles that are appearing everywhere, really are advanced robots. The tests so far tend to show that they are safer than humans already. We’re really not far from the world of I, Robot, where Detective Snooper (Will Smith) is considered crazy to use the manual mode to drive his car…
The actual problem for self-driving cars is legal issues about liabilities and all. The technology is almost ready. Volkswagen has started commercializing cruise-control, that takes the wheel on highways, for example.
Another robot I found very advanced, is the BigDog robot by Boston Dynamics, who also unveiled their Petman project this year. The BigDog robot is a legged robot that can actually walk around in the wild even in extreme conditions (in the mud, snow, or on ice).
Electric motors are not suitable for full sized legged robots. They are fragile, and not adaptable. Boston Dynamics use hydraulic pumps instead, but they are very hard to control. That’s the reason why they are the only ones to do it. But just watch a video of BigDog walking in 50cm of snow, you just won’t see that anywhere else.
In your opinion, what is the next industry or job position to be replaced with robots?
Until now, robots have replaced many jobs on the assembly line. Right now, they are used to automate even more the agriculture. In general, the keypoint is controlled environment, where humans and robots can be fully separated.
That’s why you see robots in factories: the factories are built based on efficiency, so if you need to separate robot zone and worker zone, you just do it. It’s the same in space: no humans around. In agriculture, there are not many humans either.
Autonomous tractors are being developed, but what I think is a great idea is the robotic greenhouse. Look up Harvest Automation for example. Their product is great. They create simple robots that automate key processes in greenhouses.
You have the same thing with automated warehouses: they didn’t create robots for warehouses; they created warehouses for robots, where humans have no need to go in 99% of the time.
Finally, I don’t think we’ll see robot maids anytime soon. It would be so much easier to create robotic homes instead. If you can conceive a house from the start so that robots can automate all the chores that would be possible and efficient. Think of double walls, with all the mechanics hidden from the eye. The table is against the wall, so that robots can access it easily before and after meal and all. When I have enough money, I want to work with an architect to conceive such a home. I think construction companies should start developing these kinds of houses today.
There are always movies about robots trying to take over. Do you think this is possible? How far out do you believe this could happen?
I think these kind of movies are cool. A little like the books and movies where monkeys are taking over humanity. The truth though is that monkeys are physically unable to speak, and cannot be as intelligent as humans. They just don’t have the vocal track and brain for that.
All the same, the robots cannot evolve on their own, and the decisions they take are strictly confined. What is possible is hacking a robot, turning it over its own army. But in the end, there will always be a human somewhere to give the orders.
What are your future goals?
I will graduate from my Master in robotics in October. I plan to find a job in a robotics company and create new robots. I may create my own company one day, making robotic homes or whatever idea I have.
Also, I want to expand my website. I want http://www.about-robots.com to become a reference in learning to build robots. So I will keep on working on it in my spare time.