He created a robot that plays Jenga and explains how it could transform the mobile phones of the future

Learning during childhood is a good example to confirm that the game teaches skills that go beyond pure entertainment. A researcher at the Mbadachusetts Institute of Technology also believes that some complex skills are learned by playing. To prove it, he created a robot who plays Jenga and that in the future could be an efficient operator in factories.

"We believe thatThe skills required for this game are common to many other manipulative tasks.. We are particularly interested in industrial automation, "he said in dialogue with TN Tecno Nima Fazeli, a member of MIT's engineering department and project director we are planning here.

According to reports, this robot with artificial intelligence surpbades humans in different variables of this game, but not all. At least for the moment.

– Why did you choose Jenga and not another board game for this development?

– It happens to meet some of the essential skills required in many areas of our life and work. To play effectively, the robot must combine sight and touch. He must also be able to understand the different possible interactions between his members, the blocks and the tower.

The maderitas are sometimes blocked, sometimes they move. That's why he learns that there are different types of behavior. It's not enough to see: he need to relate to the tower to learn more about blocks and their behavior. And he is obliged to learn quickly.

– The skills of this automaton could be useful in other areas and tasks beyond the game?

– We want robots to acquire skills on industrial lines with great speed, in order to meet the changing requirements of modern production lines. For example, every year phones are updated. The robots must therefore adapt quickly to the new demands of production to make these phones and be able to respond with dexterity to the mistakes they make.

– We know that, imitating the human being, the machine combines reasoning and motor skills. In this sense, is it more difficult for the robot to know which block in the tower should be moved? Or that his movements are subtle?

– Our project is focused on the robot including the mechanics of the Jenga tower. This means that you have learned how the blocks move when you touch them. The system uses these sources of information to decide whether to push and how to extract each piece of wood. The challenge was not only that he was skilled in his movements: we create a robot that thinks of strategies, instead of one that takes random blocks.

– Is he more skilled than a human being for this game?

– This machine (a robotic arm with two limbs that work like fingers) is able to measure the force much more accurately than a human and can better control their movements. Human hands are more complex.

However, in this case of development a person can beat the robot when he really tries. It turns out that the automaton does not yet have the same level of dexterity and the ability to move freely around the tower as a human.

Fazeli explains that the challenge is precisely to develop a mechanism that reaches the level of competence of a human being. "We are now working towards better models and representations. the robot can use this information to learn more about the world and plan its interaction with it"He says.

As investigations progress for the system to reason as Jenga's most skillful player, there is an aptitude in which, no doubt, he surpbades us: his mechanical fingers never tremble.