Positive Technology Journal

via KurzweilAI

University at Buffalo researchers are developing brain-computer interface (BCI) devices to mentally control robots.

“The technology has practical applications that we’re only beginning to explore,” said Thenkurussi “Kesh” Kesavadas, PhD, UB professor of mechanical and aerospace engineering and director of UB’s Virtual Reality Laboratory. “For example, it could help paraplegic patients to control assistive devices, or it could help factory workers perform advanced manufacturing tasks.”

Most BCI research has involved expensive, invasive BCI devices that are inserted into the brain, and used mostly to help disabled people.

UB research relies on a relatively inexpensive ($750), non-invasive external device (Emotiv EPOC). It reads EEG brain activity with 14 sensors and transmits the signal wirelessly to a computer, which then sends signals to the robot to control its movements.

Kesavadas recently demonstrated the technology with Pramod Chembrammel, a doctoral student in his lab.  Chembrammel trained with the instrument for a few days, then used the device to control a robotic arm.

He used the arm to insert a wood peg into a hole and rotate the peg. “It was incredible to see the robot respond to my thoughts,” Chembrammel said. “It wasn’t even that difficult to learn how to use the device.”

The video (below) shows that a simple set of instructions can be combined to execute more complex robotic actions, Kesavadas said. Such robots could be used by factory workers to perform hands-free assembly of products, or carry out tasks like drilling or welding.

The potential advantage, Kesavadas said, is that BCI-controlled devices could reduce the tedium of performing repetitious tasks and improve worker safety and productivity. The devices can also leverage the worker’s decision-making skills, such as identifying a faulty part in an automated assembly line.