Cognitive systems A baby robot to teach and learn from
Robotic dogs, cats and even dinosaurs have tickled public interest in recent years, but what about a robot baby?
The iCub is a crawling, learning android the size of a two-year-old child. Yet it is no toy intended for tech-loving consumers.
The robot is a serious endeavour by international researchers working in the ROBOT-CUB project to investigate artificial cognition – the means by which robots can be made to understand and interact autonomously with the world around them.
The ROBOT-CUB researchers’ work should lead to more advanced robots able to assist and interact with humans, while at the same time helping to deepen our understanding of the cognitive functions of the human brain.
A test bed for research
The iCub is essentially an open test bed for the ROBOT-CUB partners and other researchers around the world to experiment on. With 53 degrees of freedom of movement, the iCub is able to crawl on all fours and sit up, move its head and eyes and use its hands to grasp objects. Cameras and sensors allow it to see, hear and touch.
The robot’s software brain can be made to learn from experience – by observing others and through trial and error, much as a young child does.
The researchers are teaching it to perform complex tasks, such as tracking an object across a room, navigating based on fixed objects around it, and predicting the consequences of how it interacts with an object – skills that humans learn naturally as children.
Available for free to all
The original team that developed the iCub are not alone in using it for research, however. By allowing other researchers around the world to use the robot and to develop their own for free, the ROBOT-CUB project is helping to spur new initiatives using the iCub as a base.
A competition for ideas run by the project partners in 2007 resulted in six research labs being given an iCub on which to experiment.
The new projects range from replicating the functions of certain neurons in the human brain in digital applications to programming the iCub to link verbal commands to physical objects.
The projects are being conducted by labs in Barcelona, London, Paris, Ankara, Munich and Lyon.
iCub’s software coding, along with technical drawings, are free to anyone who wishes to use them.
The software and hardware were designed using a modular system that allows large numbers of researchers to work independently on separate aspects of the robot.
Learning about ourselves
The ROBOT-CUB partners also note that work on artificial cognition in robots could help us better understand how the human brain functions, thanks to crossover research into robotics, psychology and neuroscience in recent years.
Fields of science
Call for proposal
See other projects for this call
Funding SchemeIP - Integrated Project