It uses radio frequency identification tags to locate items and a laser range finder to avoid collisions.

It was created by professor Vladimir Kulyukin at Utah State University and shop floor trials have already begun.

Prof Kulyukin and his colleagues are in negotiations with a large supermarket chain to conduct more extensive trials.

"We refer to it as a robotic shopping assistant," he told the BBC News website.

The idea came to Prof Kulyukin after several visually impaired people told him that they had difficulty shopping independently.

"The idea is that you simply come to the grocery store, grab the shopping assistant and it leads you to the different products. When you leave the store you leave it behind."

The shopping robot has been tried out at a local grocery store - Lees Marketplace - in Logan, Utah.

"I'm a graduate of Utah State University so I like to help them in anything they do," said owner, Lee Badger.

You can also check : Top News News of the Week

"We worked out some times when we weren't open and they tested the robot early in the morning."

"As a grocer you want to attract as many people as you can, and if some day this robot could help blind people come to the shop, I want to be part of that."

Professor Kulyukin is now in talks with a supermarket chain about getting access to a larger shop for longer periods of time.

"We need to test it in a 24/7 environment," he said.

Until now, the radio tags have been placed on particular shelves in order to locate specified products, but not on everything in the shop.

The next stage will be to tag the shelves with generic categories - toothpaste or cat food for example - and then to fit the robot with a bar code reader so that it can identify every product in the shop.

First impressions

A number of visually impaired testers have given enthusiastic feedback, but not all of them were Braille users so were unable to use the Braille interface to find items.

The development team is now adding a synthetic speech option to the robot's directory.

Further work is also planned to improve the robot's motion - one woman with a back injury complained that the device accelerated too fast for her.

Although the cost of the prototype is around £8,000, Prof Kulyukin believes it could be mass produced for a third of this.

"It's still got a long way to go but it's a very useable prototype," said blind assistive technology specialist, Sachin Pavithran.

Mr Pavithran - who also works at Utah State University - has tested the robot in a number of different environments and thinks it offers the possibility of greater independence.

He says such a device would be of particular benefit to him in airports while waiting for connecting flights.

"Normally I just wait at the gate," he said.

"With a device like this I'd have the option to go to a restaurant or explore the shops."

Prof Kulyukin and his team hope that with the right investment their robot could be available to blind shoppers or travellers in a year or two's time.