’We are looking at helping people with severe movement disabilities such as motor neurone disease and spinal cord injuries,’ said Dr Girijesh Prasad, a senior lecturer at Ulster’s school of computing and intelligent systems.
’They would not be able to make any movement through the neuro-muscular medium, so first of all, we are looking at a brain computer interface (BCI) where the brainwaves would be used, often through EEG (electroencephalography) for creating an interface to command both the wheelchair and a robotic arm.
’First, we can provide the mobility through a wheelchair that has robust, inbuilt obstacle avoidance capabilities. Second, if we can provide disabled people with some sort of robotic arm, they can also attempt to do activities of daily living on their own as far as possible.’
As part of their investigations, the scientists will concentrate mainly on developing the command-and-control aspect of the intelligent wheelchair, using an off-the-shelf model donated by Invacare UK, a wheelchair manufacturer based in Bridgend, Wales.
Prasad explained that the Institute of Technology, as specialists in robotics, will work on the wheelchair and robotic arm combination, while the Ulster team will concentrate on developing the BCI and will eventually integrate the interface with the chair.
’Cameras, as well as the robotic arm, will be attached to the wheelchair. There will be a camera attached to the back of the wheelchair behind the person, so that images of the environment as well as of the robotic arm (to help it ’see’) would be obtained in quick succession, and we will look at the patterns of the images and do the image processing.
’Using that information, we can then manoeuvre the wheelchair, and would be working to develop the manoeuvring of the robotic arm,’ said Prasad.
The scientists hope to enable the disabled person to control the wheelchair through electrical signals in the brain gathered by a commercially-available electrode cap worn by the user. The data is then transmitted from the electrodes to a small pocket computer or laptop.
’In the EEG-based BCI, if a person thinks of moving, say, his or her arm to the left, we get one type of pattern in the EEG in the sensory motor cortex, and when the person thinks of moving to the right, we get a slightly different pattern,’ said Prasad.
’Through intelligent signal processing, we can find out what pattern is appearing at any one point in time and by that we find out whether the person is thinking of moving to the left or the right. With that, we can devise a system by which a sequence of left movements and right movements can be converted into a command.’
There will also be a scene reader attached to the wheelchair to display images of the environment in a comfortable position for the wheelchair user.
’Using the reader, disabled people will be able to point out their desires in terms of moving the wheelchair or the arm, by thinking about the direction in which they want to move using the BCI.’
In addition, Prasad plans to develop a virtual keyboard within the BCI to allow a severely disabled person to communicate. By just thinking, the person would be able to ’type’ the information they want to convey and a speaker would read it out.
One of the main challenges Prasad’s team will have to contend with is the noisy nature of EEG signals, which can make it more difficult for the signals to be accurately analysed.
’As well as being a noisy signal, EEG is also time varying, so it varies from movement to movement and depends upon the physical or mental state of the person using it,’ said Prasad.
’However, we have been working on BCIs for years and have developed some signal processing algorithms which we hope will provide much better robustness to the system in terms of tackling the noisy characteristic of the EEG signal.’
According to Prasad, the only existing competitor for the new wheelchair is a BCI-driven wheelchair in its preliminary stages of development in Switzerland. But the Swiss version does not have a robotic arm, and has only been tested on able-bodied people.