Advanced brain-computer interfaces that can be used to control the complex motion of powered robotic arms and hands have relied on electrodes implanted in the brain. These have serious downsides that currently make them impractical for use outside of scientific research. A team at University of Minnesota has now created a control system that relies on traditional non-invasive electroencephalography (EEG) to allow people to use a robotic arm to grasp, move, and place objects in front of them. We’ve seen attempts at this before, which were quite impressive but haven’t achieved the natural and seemingly easy movements demonstrated by the Minnesota team.
The study involved eight volunteers who learned how to move the robotic arm by simply thinking about it. This process involved different iterations at first involving moving a mouse on a screen and later moving up to complex 3D motion of the robot. Eventually they were able to stack objects on a three-level shelf with reasonable accuracy, certainly an impressive feat that gives severely disabled people a preview of what they should expect to be doing in the not too distant future.
Here’s a University of Minnesota video showing off the capabilities of the new system:
Study in Scientific Reports: Noninvasive Electroencephalogram Based Control of a Robotic Arm for Reach and Grasp Tasks…