MIT engineers have developed a robot equipped with a soft-pronged gripper, a force-sensing wrist cuff, and an external camera - features which enable it to effectively play the game, "Jenga."
This robot can (probably) beat you at Jenga
00:53 - Source: CNN Business
CNN  — 

Robots are an increasingly common feature of everyday life, whether they are cleaning your house or stacking shelves at the grocery store.

Now engineers at the Massachusetts Institute of Technology have developed a robot that performs a seemingly more frivolous task: playing Jenga.

The popular game involves removing one block at a time from a stack of 54 arranged in 18 layers of three.

Jenga requires increasing levels of concentration and dexterity to remove the blocks without knocking over the whole tower.

The MIT machine, however, appears perfectly equipped for the task.

With a soft-pronged gripper, external camera and force-sensing wrist cuff, the robot can both see and feel individual blocks in a Jenga tower, according to an MIT statement.

It “learns” whether to remove a specific block in real time, using visual and tactile feedback, in much the same way as a human player would switch blocks if the tower started to wobble.

Full details of the research were published Wednesday in the journal Science Robotics.

For Alberto Rodriguez, assistant professor in MIT’s Department of Mechanical Engineering, the robot represents an advance in the use of tactile, physical interactions to learn new tasks.

“This is very difficult to simulate, so the robot has to learn in the real world, by interacting with the real Jenga tower,” Rodriguez said in the statement.

“The key challenge is to learn from a relatively small number of experiments by exploiting common sense about objects and physics.”

And Ali Shafti, a research associate in robotics and artificial intelligence at Imperial College London, said he’s impressed by the approach. He was not involved in the MIT research.

By synthesizing what the robot needs to learn into a small number of concepts, rather than making it compute every possible variation, researchers allowed the machine to learn more efficiently, Shafti told CNN by phone.

“This is the way robotics and AI needs to be moving together,” he said, adding that the robot learns by playing in a similar way that a human would.

During a few informal trials, the robot matched up well against human Jenga players.

“We saw how many blocks a human was able to extract before the tower fell, and the difference was not that much,” study author Miquel Oller said.

However, the robot was not able to use the same strategy as a human player would in a competitive game, such as removing a block that would leave an opponent with a particularly difficult task in the next round.

While a Jenga-playing robot sounds fun, there are plenty of other practical applications for this kind of technology.

It could be used to sort recyclables from landfill trash, or assemble consumer products such as cell phones, Rodriguez said.

And Shafti said he believes the approach could be used to create robots that assist the elderly or disabled in tasks such as opening a bottle.

Developments in robotics continue to move fast, and one London airport may soon offer a robot valet service.

Later this year, Gatwick Airport will test an autonomous robot that slides a large, slender “bed” beneath vehicles and moves them to spots in the parking lot.