There aren’t many European cities that can boast with being home to a worldwide leading robotics company. We are proud – and excited! – to say that Barcelona is one of them. PAL Robotics, founded as a startup in Barcelona, builds biped humanoid robots for research and for performing services that enhance people’s quality of life. All the robots are conceived and built on site from start to finish at their office, which is constantly buzzing with robots moving around as engineers conduct demos. It’s a futuristic setting that, thanks to the people working there, still manages to remain a super friendly environment.
We visited the PAL Robotics office in Barcelona and sat down with CEO Francesco Ferro to talk about issues like how humanlike should a robot look, how robots could work together with humans, and what it’s like to be building bipeds in a world that’s not used to seeing machines walk. There is, of course, an infinite number of aspects to talk about when it comes to robotics, but let this be an introduction.
A (Very) Short History of PAL Robotics
In 2004, four engineers came together in Barcelona and built a chess playing robot that would play against the current world champion of chess. Incidentally, it was also the first autonomous humanoid biped robot in Europe, built in a mere 14 months. The company from the United Arab Emirates that had commissioned it wanted more robots, and so PAL Robotics was off to a great start.
A few successful robots later, which were meant primarily for research, PAL Robotics decided to connect business and research. In 2012, they created REEM, a wheeled robot with a touchscreen on its torso meant for acting as a service robot and providing information at shopping malls, airports or museums. The following year, they built REEM-C, a biped meant for purely educational purposes and sold to universities to study two-legged motion, perception and control.
REEM-C was followed by three more milestones, three more robust robotic platforms in just three years’ time. The more affordable and smaller sized TIAGo is one-armed mobile manipulator looks like a cute mashup between ET and Wall-E, and is capable of performing simple tasks autonomously. Stockbot is a wheeled, human-height tower built to do inventory in shops and warehouses using RFID technology. The latest, most advanced robot built by PAL Robotics is the mighty TALOS: a high-performance biped meant to work on production lines and help industry workers carry tools.
TALOS is nothing less than intimidating: he stands 1,75 m tall, weighs 95 kg, and looks like a real-life Transformer. When we arrived at the PAL Robotics office, the engineers were kind enough to unbox him, and he very politely introduced himself. After this initial surreal experience, we sat down with Francesco to talk about what I’ve just witnessed for the first in real life.
Interview with Francesco Ferro, CEO at Pal Robotics
“Up until a few years ago, the biggest enemy of robots was humankind.”
You envision a – fairly near – future where robots will work alongside people. So robots are not meant to replace humans?
We know that a lot of people think like that. It’s a big problem for robotics in general. Humans, from a social point of view, are very aggressive. When we sense some sort of danger, we instantly try to fight it, or escape from it. Up until a few years ago, the biggest enemy of robots was humankind. Because humans see robots as competitors. They’re afraid of being replaced.
Our vision is that robots are meant to improve the quality of life of people. What robots are trying to replace right now is automation, things that don’t carry any added value for the community. In my opinion, nowadays, people are stealing jobs from robots. I’ve seen a lot of factories where humans are the robots. They’re doing a job just because they need to work, not because it creates any added value. In some production lines people do physical work that causes severe health problems, and ultimately that’s damaging for both the workers and the companies.
What’s your position on the uncanny valley (the dip in a human’s emotional response when encountering something that’s strangely humanlike)? How much do robots need to resemble humans?
We want people to recognize a robot when they see one, and not to confuse it with a human. However, robots do have to resemble humans to some extent because research shows that it’s much easier for us to interact with humanoids. Aesthetics are very important for HMI (human machine interaction). A robot with human semblance, with anthropomorphic features, is recognized as a humanoid. Humans prefer talking to a humanoid form to something that looks like a washing machine. So we try to stay in that range: resembling a human but not too humanlike.
“No two robots are exactly the same. They all have their own peculiarities.”
Do you find it hard to send off a robot when you’re shipping it to a client? Are they a bit like your “children”?
I wouldn’t know, because I don’t have children, I have cats! (Laughs.) I wouldn’t say that they’re like my children, but what I can say is that they’re all a bit different. For example, we have several units of TIAGo, and even though we do our best to make all our robots equal, no two are exactly the same. Now I understand why car manufacturers say that all their cars may look the same, but they actually aren’t. All our robots have their own peculiarities: some are a bit more tricky on one part or another.
When we ship a robot, I don’t find it hard to say goodbye to them at all. On the contrary. I’m very happy because someone else is going to be using them and helping us reach our mission. We need collaboration, we need different use cases to provide people with.
How do you come up with use cases? Do you create the demand or is it coming from the clients?
It depends on the robotic solution. With bipeds, we are creating the market. When we started 15 years ago, everyone said we were crazy, that there would be no demand for it. And now we’re selling biped robots, so we have created the market. One way to justify our obsession with bipeds would be something that Ford said: “If you asked people what locomotion machine they wanted, they would have said faster horses that eat less.” Sometimes, you have to move away from your comfort zone and create a revolution to solve an ancient need.
With other projects, we just try to find a solution for a problem that a client has. For example, the autonomous inventory system is a response to a client’s need.
Another way to create the market is to make robots that are cheap and not useful, but people purchase them just because they can afford them. It doesn’t matter if it works or not. It just looks like it works. Most cleaning robots are like that. You buy them, and you realize that they don’t help much. They create a commodity that you don’t need, it’s not very useful, but you purchase it because you can afford it. Now this is something that we don’t want to do. We try to minimize the cost but at the same time we want to make robots that are useful. And this is the long-term vision.
“We’re one of the few companies that have robots for real use cases.”
What’s your biggest motivation?
Everyone is talking about robots now, but there are very few companies that are actually trying to bridge the gap between research at big universities and the market. A few weeks ago, we went to the European Robotics Week in Brussels, one of the biggest robotics events in Europe. We took three robots there, and there were probably another two from other companies. Nobody else had actual robots to show. So what motivates me is seeing that we’re one of the few companies that have robots for real use cases.
I also have a more romantic motivation. It has to do with healthcare. Healthcare is a field that could be very important for robotics, and vice versa. Our society is getting older and older. In a couple of years, we won’t have enough nurses to take care of all the elderly people. And, for example, with our robot TIAGo, we can demonstrate that robots could help elderly people stay in their own homes for at least two years longer than they are able to right now. We proved this at a competition that we just hosted in our office, the European Robotics League Tournament. It was a service robots contest where robots had to prove themselves useful helping an older woman in a home environment autonomously.
In order to help the elderly, robots would have to perform very simple tasks and do some regular checks. This would not only reduce the cost of social security – that’s the vision of the government – but it’s also a question of people’s dignity. For instance, I personally would love to stay at home when I get old. I think that people who are forced to live in hospitals completely lose their motivation to stay alive. I think it’s really important that we help them because the elderly have a lot of wisdom and experience that the world needs.
How long before robots become a part of everyday lives, helping us with our daily tasks?
I don’t think that we’re very far from it at all. We already use robots everyday: they’re called washing machines, cleaning robots, and so on. Of course they’re not humanoids, which is probably what your question is about. I can’t give you an exact date, but I think we’re really close to having humanoids help us out in our everyday lives. I’d say more or less ten years. I’m sure that I will live to see it.
To sum up, here are a few things I learned during my visit to the PAL Robotics office:
It’s really difficult for a robot to walk on two feet, because every time they take a step, they’re essentially preventing a fall. Also, robots have no idea of their own dimensions, something that’s super natural to humans, so they need a bunch of really expensive sensors in order to be able to dodge obstacles and not cause accidents. Thirdly, for now, humanoid robots and AI don’t really go hand-in-hand: when that happens, that’s when we’ll start to see walking and talking robots out of Star Wars in real life. And most importantly: robots are not, or should not, be made to replace humans, but to help make their lives easier.