Using their thoughts alone, five people have successfully maneuvered a flying quadcopter through a 3-dimensional obstacle course. Further developments in this field of brain-computer interfaces may lead to mind-controlled wheelchairs, prosthetic arms, and exoskeleton suits. Popular Mechanics reports.
Previous brain-computer interfaces have allowed users to interact virtually or issue commands to robots. This new system adds enhanced control, allowing users to make real-time decisions, changing course midstream.
"Eventually, this could be applied in clinical work to help paralyzed or incapacitated patients to be able to have a normal daily function of getting out of bed and moving around," said lead author Karl LaFleur of University of Minnesota. "It would provide some of the autonomy and independence that they're robbed of when they contract some sort of neurodegenerative disease."
- Five participants wore noninvasive electroencephalogram (EEG) caps fitted with 64 electrodes.
- They carried out 2D virtual tasks, like moving a ball on a screen, and practiced moving a virtual helicopter with their minds.
- They learned to associate certain thoughts with quadcopter movements: imagine clenching the right fist to turn right, imagine clenching both fists to move up. (The idea is to use signals someone paralyzed can give.)
- The EEG picked up on the electric current produced by neurons involved in imagining those tasks, then sent those signals to a computer.
- The computer translated them into a command and beamed those orders wirelessly to the quadcopter.
- A camera on the front of the quadcopter lets users see through the machine's eyes.
The subjects watched the machine's movements on a computer screen and used their thoughts to instruct it to turn, fall, and climb to navigate through a series of foam rings. One of the participants achieved 90.5 percent accuracy. The group average was 66.3 percent.
Researchers are working on refining accuracy and control, and even incorporating other body parts (feet and mouths) and smaller movements (like a finger flick).
The work was published in the Journal of Neural Engineering last week.
Image: University of Minnesota