The ideal future is that where robots become an extension of humans, and they start doing the dirty work. We are not there yet as our communication with robots depends on keyboards or remote controls that delay the information to reach the receiver. However, things are about to change in light of a recent discovery. A new interface that is connected to the human brain can control robots and allows people to correct the mistakes robots make.
Daniela Rus was one of the members of a team of scientists from the Massachusetts Institute of Technology that came up with something groundbreaking in the computer’s world. They managed to create an interface based on electroencephalography or EEG for short. This board can record the electrical activity of the brain by attaching electrodes on the scalp. Rus described the recent endeavors as a series of baby steps towards reaching the point where robots adjust automatically to our thoughts.
There were five volunteers who had to watch how a certain robot reaches out to one of the two LED lights in front of it. During the test, participants had to wear the EEG helmet while scientists measured the recorded reactions. There were several sessions in which one of the two LEDs was designated as the correct one in a chaotic way. When the robot showed the intention to reach for the wrong light, the volunteers were simply thinking of it as a wrong action. The EEG read this reaction and sent a message to the robot. The robot adjusted its actions accordingly and decided to touch the right LED light in the end.
The robot was called Baxter and was the creation of the Rethink Robots in Boston, Massachusetts. The way this works is based on error potentials. These appear in the EGG whenever a person hooked on this headset is of the opinion that an action, event, or thing is wrong. These signals are strong and have a shape of their own which enables robots to sense them immediately.
The experiment had a successful rate of 70% when people were able to control robots. The same results were recorded in a different setting as well. Baxter received the task to sort wires and paint bottles in a box. Whenever they mistook one thing for the other, the witnesses thought of it as a wrong action and sent a signal to the robot. Baxter corrected itself in 70% of the cases.
Image source: 1