A few years ago, seeing a robot performing simple tasks was a scientific breakthrough. Now science is making a new breakthrough that is close to science fiction, as scientists have succeeded in developing an artificial intelligence system that can predict human intentions before they implement them.

The new innovation was recently published in the journal “Robotics and Computer-Integrated Manufacturing” by researchers from Loughborough University in the United Kingdom, where scientists have developed an artificial intelligence system that can tell the direction of arm movement before proceeding to it.

Scientists consider cooperation between humans and robots an exciting opportunity for future industrialization, because it combines the best of both worlds; The relationship requires close interaction between humans and robots, which can benefit greatly from the robot’s anticipation of human movement.

Would you like a Robot to take your place at home?

electroencephalogram

“Ideally, for effective teamwork, ‘human and robot’ must understand each other, which is difficult due to being completely different and ‘speaking’ in two different languages,” says Achim Borkel, the study’s lead researcher.

To get around this difficulty, Burkell and his colleagues tried to give the robot the ability to “read” the intentions of its human partners by linking the activity of the human brain’s frontal lobe to the AI ​​system.

By means of this connection, every movement of the human body in the brain is analyzed and evaluated before it is carried out. This analysis can help to communicate the “transition intent” to the robot, however, brains are very complex organs, and detecting the pre-movement signal is difficult.

Loughborough University researchers tackled this challenge by training an AI system to recognize pre-movement patterns from electroencephalograms (EEGs) — tiny metal discs (electrodes) attached to the scalp that allow recording of human brain activity.

motion prediction

The study refers to the results of a test conducted with eight participants, in which they had to sit in front of a computer that randomly generates a letter from A to Z on the screen, and presses the key corresponding to the letter on the keyboard.

The AI ​​system had to predict which arm the participants would move on from the EEG data and this intention was confirmed by motion sensors.

Experimental data shows that the AI ​​system can detect if a human is about to move their arm up to 513 milliseconds before they move and, on average, about 300 milliseconds before actual execution.

In a simulation, researchers tested the time advantage effect of a human-robot collaborative scenario, and found that they could achieve higher productivity for the same task using this technology, with task completion time 8 to 11 percent faster.

Change the stereotype

Achim Burkel plans to build on this research, and eventually hopes to create a system that can predict where to direct motion—for example, reaching for a screwdriver or picking a new workpiece.

He expresses the hopes for the study, saying, “We hope that this study will achieve two things: First, we hope that this proposed technology will help achieve close collaboration between humans and robots, which still requires a great deal of research and engineering work.”

Second, the researcher and his team hope to change the prevailing stereotype that “robots, artificial intelligence, and machine learning pose a threat to the future of human labor in manufacturing.” Borkel sees this technological development as “an opportunity to preserve humans as an essential part of the future.”