Science is advancing at a breakneck pace, with all sorts of advanced technology available today previously relegated to the realm of science fiction. Recently, we took another step forward into the world of tomorrow, as MIT researchers have now devised a system that will allow robots to follow commands from humans based on thoughts.
Science News reports on this new breakthrough in robotics. According to Daniela Rus and her team (the creators), their technology has allowed robots in the lab to follow commands from humans based on just brainwaves and small hand gestures, effectively giving people the ability to direct and control them with minimal interface.
Tests were conducted on seven volunteers. In these tests, robots equipped with drills would have to pinpoint one of three targets indicated by colored lights and touch them with their drills. On their own, the robots correctly identified their targets around 700 times in the 1,000 tests. However, if they were headed towards the wrong spot, the human volunteer (wearing a special neural interface on their head) would need to think for the machine to stop.
The act of thinking was enough to make the machine stop in its tracks, after which a simple swiping gesture to the left or right would set the robot on the path towards the correct target. By doing this, the robots were able to successfully hit the mark a total of 97% of the time during testing, all but eliminating error while performing tasks more precisely than humans would reasonably be able to.
Set to debut at June 28th’s Robotics: Science and Systems conference in Pittsburgh, successful tests could revolutionize robotic workforce participation the world over in a multitude of sectors. Surgeries could be performed by robots while overseen by doctors, packages could be sorted and shipped with minimal human interaction needed, all in environments too chaotic or distracting for normal forms of communication between workers. The possibilities are nearly limitless.
Rus says that she and her team are working on better versions of this technology already, hoping to make robots that can recognize and act on a wider variety of gestures. That way, explains study coauthor Joseph DelPreto, they would more easily be able to interpret what commands a human is giving them in a more fluid manner.
Should things play out as hoped, this technology could even find application in things like self-driving cars with the system’s ability to correct its own errors at such a quick pace. In the end, the only real barrier to more widespread usage seems to be with how to streamline the uncomfortable headsets.
