For several years now, leading world experts have been developing in the field of creating neural interfaces. And recently experts from the Artificial Intelligence Laboratory of the Massachusetts Institute of Technology (MIT) succeeded in this, who were able to control the robot using brain impulses, and for this they used quite affordable technologies.
The new system builds on a previous MIT development. However, it had a significant drawback: to control something, the operator would have to "think in a special way." The human needed to look at certain displays or interface elements that corresponded to certain functions of the robot. Moreover, the system required calibration for each specific person, and in general, such management made life more difficult than easier.
For the new robot control system, MIT engineers used the well-known electromyography (EMG) in conjunction with electroencephalography (EEG). Now, for control, not only brain waves (as it was before) are recorded, but also muscle movements. As study author Joseph Del Preto stated, “This made the management process much easier. Combining EMG and EEG made interaction with the robot more intuitive. By fixing both muscle movements and brain waves, we more easily transfer human manipulations to control. This helps to facilitate communication with the robot and make it similar to interaction with a human."
Remarkably, to implement the idea, scientists used the so-called signals called "error potentials" (ErrPs). In the course of work, they were generated every time people gave the robot one command, and he performed another. At the same time, the person, as it were, "thought about how to execute the command correctly." Thanks to this, an algorithm was developed that is based on thoughts about the manipulation of the robot, and not on "looking at different displays." You can learn more about the development and management process in the video below.
Vladimir Kuznetsov