MIT brain robot interface

Scientists have studied ways to interface a human brain with a robotic device. One goal is to let a person control a robotic prosthetic limb using his or her mind. It’s a complex problem that is difficult to solve.

Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and Boston University have taken a different approach. Rather than have the person control every movement of the robotic device, they have instilled intelligence in the robot so that it can undertake tasks autonomously. The person observes and then corrects the robot if it appears to be making a mistake. The brainwave analysis can detect these error signals in just 10 to 30 milliseconds. The subject doesn’t have to think of the correct motion, or try to control the device; all the person does is think “That’s wrong,” and the system can detect it. The robot can then try a different action.

This technology could have enormous benefits in industrial settings, where a human-robot partnership could be much more productive and accurate than either could be on their own. But it also has great potential to help people with various impairments. A user could issue a voice command to the intelligent device (which could be a prosthetic or a stand-alone unit), and then observe to correct the robot if it does not execute the instruction correctly. This could provide a far more efficient and effective interface for users, and increase their independence.