By combining artificial intelligence and robotics, Swiss researchers have created a mechanical hand capable of learning human movements by imitating those it sees.
And if the era of humanoid robots was closer than ever? Robotics has largely benefited from the dazzling progress ofartificial intelligence, so much so that for some, AI will reach its next stage of evolution once placed in a body. In the meantime, much research is focused on reproducing the most complex human abilities: taste with an artificial tongue,smell with an AI capable of recognizing good wines… But without talking about these very specific faculties, there are some that have been a problem for a long time.
If robots are capable of washing dishes for example, most operations requiring a fine motor skills are still an obstacle. We are thinking in particular of the seizure of fragile or small objects, or even seemingly simple operations, but requiring a very precise sequence of movements that must be reproduced without making a mistake. It is with these limitations in mind that Swiss researchers created a universal robot hand.
This robotic hand learns by imitating human gestures
Rather than seeking to build a robot focused on a particular function, they chose to build a hand capable of attaching to any existing robotic arm. To facilitate learning, the integrated AI allows it to learn movements by imitating those it “sees” on videos featuring humans. There are multiple applications: in industrial bakeries to gather and package bread or pastries, in pharmaceutical laboratories to run diagnostic machines, etc.
Read also – This humanoid robot has made immense progress in just one month, we are at the dawn of a revolution
“Our vision is to have a basic, general-purpose model so you can give commands by talking to the robot, like “Robot, pour me a beer »“, summarizes Elvis Nava, co-founder of the startup mimic. It can also make coffee as seen in the illustrative image of this article. Several companies are already interested in purchasing the robotic hand.
Source: Sifted