A dialogue with Giorgio Metta
2 June 2016The Dialogues allows participants to meet with personalities from the world of business and culture that have achieved excellence in their respective fields. They are an opportunity to learn about their growth path and the challenges they have faced. We have talked to Giorgio Metta, scientific vice director of the Istituto Italiano di Tecnologia and creator of the iCub. Guest speaker at our Innovation Talks Spring Edition.
Giorgio Metta, an electronic engineer, is the deputy scientific director and director of the department developing the iCub robotics project of the Istituto Italiano di Tecnologia in Genoa (IIT). We interviewed him on the occasion of the Innovation Talk to which he’d been invited as keynote speaker: we asked him about the innovations concerning the development of a new humanoid robot, designed to enter the consumer market.
Mister Metta, what are the objectives of the Istituto Italiano di Tecnologia?
The IIT’s mission is to do research and technology transfer. We started thinking to shift research towards a production area. We established ourselves in the field of research by publishing articles aimed at attracting additional funding. Since 2011 we’ve started investing more on technology transfer: creating startups and directing ideas, developed thanks to research, towards the market, especially by selling patents and creating startups. As of now, the Institute has developed about of dozen, some with local sponsors, others abroad (one in Switzerland, one in the USA), even though we prefer fostering the creation of startups here in Italy.
And as for robotics, the crown jewel of your research?
As for robotics, we’re trying to optimise the 70 million euro investment in IT for the creation of technological products that are complete, leaving aside the production of individual technological parts. This is possible thanks to the change to the Italian law that prevented a public entity from participating in a company. Therefore, we took advantage of the circumstance to create a robotics project aimed at a large market, that costs thirty times less than the iCub, our experimental prototype. We have potential sponsors with whom we’re negotiating, we have 5 million euro we received for the development of industrial projects and we’re focusing on this new robot, which is a business project, and no scientific publications are foreseen for this.
How are you proceeding?
We prepare a business plan, we posit scenarios that are functional to the launch of a robotics startup. In parallel, there’s another activity, still on robotics, financed through a joint lab with INAIL [Italian national institute for the prevention of accidents at work] on motor rehabilitation. We developed and industrialized machines for rehabilitation, some are already being clinically tested in a few hospitals in Liguria. We’re developing an innovative hand prosthesis, and the design stage is rather advanced, as well as exoskeletons to help people, who have serious difficulties moving, walk more easily. We also have a licence patent for an automotive company that is working at the automation of interiors using the sensors that cover the surface of the robot.
Let’s go back to your robot for all budgets.
For the moment, we only confirmed the shape of the robot, its provisional name is R1. The first prototype has been created in April. In order to reduce costs as much as possible, we replaced the metal parts with polymeric materials, like Teflon or variations of Teflon, sturdy materials. We used low-cost motors, replacing the aeroplane engine we’d used for the iCub prototype. This money-saving activity has lead, in terms of additional result, to the identification of solutions that have become innovative patents. And all of this for a final cost of the product amounting to about 10,000 euro.
How does the R1 robot learn?
A few skills are learnt through direct experience and others are pre-set. One of the things the machine can’t do by itself, for instance, is recognising objects on the spot. This happens through a training stage during which objects are shown to the robot, and then the robot memorizes, processes and records them using cameras. Other pre-set capacities are deployed by loading specific apps, like for example, the operations needed to prepare a coffee.
Are there competitors on the market?
There’s only another similar robot, produced by a Japanese company. It’s a low-cost product, that’s already sold 7000 units. But this robot can’t handle much. While R1 has got hands: it can open cabinets, drawers, pick up objects and hand them. It has a sensor to control the strength, in order to prevent it from breaking the objects it encounters and avoid injuring people. In addition, it gives in the presence of strength applied by others: if I push it, it doesn’t respond applying strength.
The production time-frame?
Investors would like to see the first version within 15/18 months, in order to organise demo installations or real-life tests to define further changes. We’re talking of quite a short time-frame, especially because the Japanese are already selling their product. Over this period of time, we need to fine-tune a product that certainly is not an information kiosk on wheels. Our robot will be used for physical assistance, especially for elderly people with reduced mobility. It’ll have to take care of human beings and the house, as it is at the same a communication centre and a surveillance system. All of this in about 130 centimetres of height, that could be extended by further 15 cm, with arms 60 cm long, and they too can be extended.
What do market research activities say?
We conducted one, commissioned by a company that could be interested in working with us. It turned out that the reference market is mostly made up of dependent elderly people. The figures concern 70 million elderly just in the States by 2030, that is 20% of the entire population. In China, another important market, the figure rises to about 200 million potential buyers accounting for 15% of rich or well-off elderly. Furthermore, the research showed a future decrease of night and surveillance services for hospitals. Therefore, we estimated, for the first year, a sale of 3000 units just for the Chinese market. The expected revenues should be between 50 and 100 million euro per year.
What could you reply to those who fear a future invaded by robots?
A machine can read 30,000 articles in a very short period of time but can’t create logical and creative links, that only humans can do. And this is an imponderable element linked to intuition and to such subtle processes that we’re far away from reproducing in a robot. As for the risk of men being replaced by machines in working activities, I’d rather say that it’ll solve the problem of the need of labour force, which will decrease with the ageing population. So, rather than competing, robotics could cooperate, replacing those jobs that people no longer want to perform or that are risky for humans. Let’s say that in the development of robots, the major risk I see is the persistent lack of Italian sponsors. And this is a real concern for the technological future of our Country.
