A TEAM of scientists from the North-East has made an important breakthrough in developing artificial intelligence, with a system that understands language and obeys instructions.

The experts from the University of Sunderland have won a national award for their innovation, which imitates the way the human brain works.

The team beat off competition from the UK and abroad to win the British Computer Society's (BSC) Machine Intelligence Award, which recognises developments in computational intelligence.

The academics impressed judges with a demonstration of a visually and verbally guided grasping robot at the BSC conference in Cambridge. Strict guidelines were set down to establish to what extent the robots represented an advance in artificial intelligence.

The team, led by Professor Stefan Wermter, focused on the practical use of visual recognition and navigation, generally difficult when working on machine intelligence.

Innovative techniques from neural networks - a network of electronic components designed to mimic the human brain - were integrated into the Mirror-neuron Robot Agent (MIRA) robot.

Professor Wermter said: "We are extremely pleased to have won this award. It is recognition for our research, the field of intelligent systems and for the University of Sunderland.

"One novel aspect was that the robot was controlled by artificial computational neural networks. The robot had been trained through the use of neural networks to approach and grasp an object.

"We are working on the next generation of intelligent robots, and we believe that learning and neural networks will be a key in making them smarter and more robust."

The winning team included Dr Cornelius Weber, Chris Rowan, Alex Zochios and Mark Elshaw.

The university's School of Computing and Technology set up one of the first masters in artificial intelligence last year.