Robotic wheelchair controlled by thought alone
| Robots et médecine |
(New Scientist) PARALYSED people could navigate around buildings using a robotic wheelchair that can be controlled by thoughts alone. To direct the wheelchair, users concentrate on one of several possible destinations in a 3D map of their surroundings, which appears on a screen in front of them. Users wear a skullcap with 16 electrodes positioned around their scalp. These detect millisecond-long fluctuations in brain activity, which happen when the user looks at the point on the screen that coincides with a target destination.
As the user scans the on-screen map a blue spot flashes in sequence over each possible destination. When it flashes over the point the user is looking at, it provokes a fluctuation in their brain activity that is detected by the electrodes. The computer compares the sequence of flashes with the user's brain activity, and when a flash coincides with a fluctuation, it interprets that as indicating where the user wants to go.
In 2006 the team demonstrated that four children with cerebral palsy could use a touch screen built into an earlier version of their wheelchair, to navigate around busy and constantly changing environments during their school day.
In a study to be presented at the International Robotics and Automation Conference (ICRA) later this month, they tested how well five healthy volunteers were able to control the wheelchair with their thoughts alone.
Since the pattern of electrical activity in the brain of any individual is unique, each volunteer trained the system to detect their responses by concentrating on various points in a series of test images. They then learned how to control the wheelchair by steering it around a virtual circuit.
Once the volunteers had completed this test they were let loose in the real wheelchair, and asked to complete two circuits around the lab - one designed to test how well the chair could negotiate obstacles and one to demonstrate navigation in an open space.
Minguez says the volunteers learned to use the wheelchair confidently in about 45 minutes. "The purpose of this work was to demonstrate the usability of the wheelchair," he says. "All the subjects successfully solved all the navigation tasks and learned how to deal with the device in a similar way," he says.
It's not the first wheelchair to be controlled by brain waves, but it is the only design to incorporate mind-control in a system capable of real-time navigation, route planning and collision avoidance, says Palaniappan Ramaswamy from the University of Essex in Colchester, UK.
It is the first wheelchair to incorporate mind control in a system capable of navigating in real time
The prototype chair can so far only process two thought commands per minute. However, since the wheelchair navigates itself from point A to point B automatically, the command rate is less important than for systems where the user has to instruct the wheelchair at every step along the route. Automatic navigation also reduces the mental burden on the user, says Minguez.
So far, volunteers have only been in the chair for a maximum of 2 hours. Ramaswamy points out that if people want to use the chair for longer, as they would in real life, you would need to find a new way to attach the electrodes to the scalp. The wet gel used now begins to dry up after a couple of hours, reducing the system's performance.
Minguez says he is aware of this problem, and other research teams are already working on alternative ways to attach the electrodes. The Zaragoza team is trying to increase the command response rate of their chair without reducing its accuracy. They also plan to give users more control by allowing them to issue commands to the wheelchair once it is moving towards its destination, allowing them to change course while on the go.
Once the volunteers had completed this test they were let loose in the real wheelchair, and asked to complete two circuits around the lab - one designed to test how well the chair could negotiate obstacles and one to demonstrate navigation in an open space.
Minguez says the volunteers learned to use the wheelchair confidently in about 45 minutes. "The purpose of this work was to demonstrate the usability of the wheelchair," he says. "All the subjects successfully solved all the navigation tasks and learned how to deal with the device in a similar way," he says.
It's not the first wheelchair to be controlled by brain waves, but it is the only design to incorporate mind-control in a system capable of real-time navigation, route planning and collision avoidance, says Palaniappan Ramaswamy from the University of Essex in Colchester, UK.
It is the first wheelchair to incorporate mind control in a system capable of navigating in real time
The prototype chair can so far only process two thought commands per minute. However, since the wheelchair navigates itself from point A to point B automatically, the command rate is less important than for systems where the user has to instruct the wheelchair at every step along the route. Automatic navigation also reduces the mental burden on the user, says Minguez.
So far, volunteers have only been in the chair for a maximum of 2 hours. Ramaswamy points out that if people want to use the chair for longer, as they would in real life, you would need to find a new way to attach the electrodes to the scalp. The wet gel used now begins to dry up after a couple of hours, reducing the system's performance.
Minguez says he is aware of this problem, and other research teams are already working on alternative ways to attach the electrodes. The Zaragoza team is trying to increase the command response rate of their chair without reducing its accuracy. They also plan to give users more control by allowing them to issue commands to the wheelchair once it is moving towards its destination, allowing them to change course while on the go.