There have been various ways to control a robot. You can use a remote control, write a code, and even train them to teach themselves to learn new things using machine learning. However, a brain-controlled robot sounds quite interesting if you’re too lazy to use your hand to control them.
Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and Boston University have released a study about a feedback system, which read’s human brain responses while monitoring a robot at work. Named Baxter, the robot is sent a signal to correct its error if you can recognize it making one.
“Imagine being able to instantaneously tell a robot to do a certain action, without needing to type a command, push a button or even say a word,” CSAIL Director Daniela Rus said in a statement. “A streamlined approach like that would improve our abilities to supervise factory robots, driverless cars, and other technologies we haven’t even invented yet.”
To make this brain-controlled robot listen to you, you need to wear an EEG cap. The CSAIL’s system requires reading and recording your brain’s activity. The system created a machine learning algorithm that classifies brain waves within 10 to 30 milliseconds, focusing on finding error-related potentials, also known as “ErrPs” and act on them.
“As you watch the robot, all you have to do is mentally agree or disagree with what it is doing. You don’t have to train yourself to think in a certain way — the machine adapts to you, and not the other way around,” explained Rus.
As you see in the video, Baxter is trying to recognize spray cans and wires to put them in the right box. If the operator thinks the robot is making a mistake, it immediately corrects itself as it can detect what its operator is thinking. And the best part is the operator doesn’t even need to talk or make a body gesture.
The brain-controlled robot has a lot of potential use. Researchers are hoping to use this method for self-driving cars, to supervise factory robots, etc. Although the robots are highly programmed to perform specific tasks, error-potential systems could help us train them in a way, which comes more naturally to humans.