The MIT Artificial Intelligence Laboratory (CSAIL), along with a team from Boston University, is working on this development, creating a feedback system that allows humans to correct a robot’s mistakes from the mind. To control a robot with the mind, the human brain emits a signal called Error Potential that appears at the moment in which the person perceives a possible failure. Taking this warning as a reference, the machine will know whether or not the human being agrees with what he/she is going to perform and will modify his/her behavior accordingly. Using the data of an electroencephalography (EEG), the monitor that records the brain activity detects if a person observes that the robot is going to commit a failure. To reach this point, the system itself has learning algorithms that allow to recognize and classify the different brain waves in a matter of thirty milliseconds.
As you can see in the video, to test the method of controlling a robot with the mind they used Baxter from Rethink Robotics, the company led by former director and co-founder of iRobot CSAIL Rodney Brooks. The machine had to deposit paint cans and cables in the corresponding containers. “Imagine that you can order something from a robot and do it instantly, without typing a command, pushing a button or saying anything. Everything the robot has to do is mentally linked to what the robot is doing. A person does not have to train to think in a certain way. The machine adapts to you and not the other way around.” explains Daniela Rus, the current director of CSAIL. Rus and his team believe that this system that allows controlling the robots with the mind will give a greater capacity of supervision of other robotized products, autonomous cars and various technologies that are to be invented. So, what do you think about this new technology? Simply share your views and thoughts in the comment section below.
