Robot see, robot do: Inside the autonomous android lab

This article was taken from the February 2012 issue of Wired magazine. Be the first to read Wired's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online.

On the outskirts of Genoa, the Italian Institute of Technology (IIT) is home to numerous disciplines geared toward bringing artificial intelligence into the real world. A research skunkworks with an annual budget of ¤100 million (£87m), the IIT explores robot forms of every kind: bipeds, quadrupeds, torso-less legs capable of gymnastics, and iCub (left), a baby-faced humanoid which is central to the institute's studies. Researchers have taught it to crawl, recognise objects, pick them up, move them -- and not to break them. "To live among us, robots will have to be very different from how they are now," says Roberto Cingolani (above), the institute's director. Here are just a few of the scientists involved, and how their specialist fields will take robotics out of the lab and into our lives.

Name: Roberto Cingolani

Job: Director, Italian Institute of Technology

Project: iCub

Function: iCub is an open-source humanoid robot testbed for artificial-intelligence research. More than 20 laboratories worldwide are using an iCub for robotic cognition research.

Name: Darwin Caldwell

Job: Director, Advanced Robotics Department

Project: "Compliannt" Coman

Function: The gait of these walking robot legs emulates that of a human. Including tilting the pelvis between steps to in crease stability.

Name: Sylvain Calinon

Job title: Team leader, Advanced Robotics Department

Project: Learning by imitation

Function: The software that controls this robot arm has learned to roll dough for pizza by copying human movement. iCub learns in a similar way.

Name: Giorgio Metta

Job title: Senior researcher in Robotics, Brain and Cognitive-Sciences Department

Project: iCub

Function: iCub's hand has 48 tactile sensors in the palm, and 12 in each fingertip. This allows it to pick up fragile and irregularly shaped objects.

Name: Giulio Sandin

Job title: Director, Robotics, Brain and Cognitive-Sciences Department

Project: iCub

Function: Using tools such as algorithms and sensors, iCub learns by exploring the world around itself.

Name: Chiara Bartolozzi

Job title: Post-doctoral researcher in Robotics, Brain and Cognitive-Sciences Department

Project: iCub

Function: Its head components and vision systems enable iCub to see in stereo 3D and identify numerous objects.

This article was originally published by WIRED UK