It will not look like a standard robot - it will be desk-based with a robotic arm and cameras for eyes Prof Mark Lee

Aberystwyth University academics are working on the machine which they hope will recognise objects and retrieve them using an arm and cameras for eyes.

However, it is unclear at present what it could be used for ultimately.

Scientists said they aimed to "unravel" how a part of the brain worked and would then use that information to develop the machine.

Led by Professor Mark Lee, the team from Aberystwyth is joined by academics from six other universities on the five-year project.

Financial backing worth some £1.9m has come from the Engineering and Physical Sciences Research Council - Aberystwyth's share is £470,000.

Most robots are programmed to make certain decisions, but are unable to think for themselves.

According to Prof Lee, the project's purpose was to try to "unravel" the way in which the brain worked and to then build a robot that could "think" for itself.

Prof Lee will be working with other universities on the project

"Humans and animals adapt their actions according to what surrounds them, and are able to do several things at the same time and learn from their mistakes," he said.

"With this project we hope to solve this problem of multi-tasking by using our knowledge of how the brain works."

The robot would also be able to detect items, assess their significance and prioritise, focusing on the most important.

However, Prof Lee said that it would not look like a standard robot but be desk based with a robotic arm and cameras for eyes.

"All these capabilities will be combined within an overall control system that makes use of a central selection mechanism, just as we believe occurs in the brain," he added.

"Our understanding of how the brain works is also key to the next stage which will involve teaching the robot how to react to things that change around it.

"For example, something which could potentially distract it from the task it has been set. It will also be able to learn from its mistakes just as humans do.

"Once the robot has been constructed we will then stand back and ask the question 'what general features of the model gave it its ability to integrate its behaviours successfully?'

"By doing this we hope to be able to transfer our work into a wider range of robots designed for many different tasks."