Brain impulses control computer

A researcher working at the University of Rochester, New York has developed a way of letting people control a computer using impulses in the brain.

Graduate student Jessica Bayliss has developed a way of filtering out background electrical noise from the neural signals that can be used to tell a computer what to do. Volunteers in the lab wearing a virtual reality helmet and fitted with electrodes to pick up the neural impulses can control elements in a virtual world, including turning lights on and off and bringing a mock-up of a car to a stop.

Bayliss has homed in on a particular type of signal, known as the 'P300 evoked potential'. It is generated towards the top of the back of the head and is effectively a type of recognition signal.

'We listen for when several neurons suddenly say 'that's it!" said Bayliss. She looks for this signal to occur in sync with a light flashing on the TV or stereo. If the rhythm matches the blinks of the stereo light, for example, the computer knows the person is concentrating on the stereo and turns it on.

A person does not have to look directly at the stereo. As long as the object is in the field of view, it can be controlled by the person's brain signals. As it is not necessary to move the eyes, this system could work for paralysis patients who are completely 'locked in', a state where even eye blinks or movement are impossible.

A PC carries out the processing. The software filters out background noise from the P300 signal then tries to fit that with an event in the virtual world.

Bayliss said: "We're using a robust Kalman filter. It's basically a fancy way of doing template matching. We are looking for a specific bump and we know what it looks like. Right now, we are using some regression pre-processing to get rid of extraneous signals such as eye movement."

She aims to make the source code available so other researchers working in the area of direct neural control can use it.

RoboShop

Siemens has launched a range of intelligent robots for cleaning commercial locations such as supermarkets. They are based on a Siemens' navigation system, which allows them to operate independently and orient themselves automatically.

The robot learns its environment by being led around its surroundings and shown a cleaning path. Its 'eyes - an integral laser scanner - build a map of the area, so the robot is then able to guide itself with the aid of landmarks.

The robots are made by Hefter Cleantech and were jointly developed with Dutch cleaning specialist RTB.

RoboCup

For everything football, from table-top dribblers to fully autonomous anthropomorphic soccer players, visit the Robot World Cup Initiative's web site...

See what the German robot football team team are up to...