There are traditions that are no joke. This is the case of the Swiss raclette, a subject on which doctoral students in robotics have looked to design the “Roboclette”, a robot capable of serving the raclette, quite simply.
No offense to France, Switzerland is also the country of cheese. Among the traditional dishes that are served to warm up, we find raclette. And when the technology gets tangled up, the result is a robot that took a lot of work to be able to serve raclette cheese the way it should be.
When tradition meets innovation
The, the ?) Roboclette is a robot equipped with two articulated arms, each of which has a specific mission: the first holds the half wheel of cheese, the second scrapes the cheese for serving on the plate. The handling looks simple, but it’s actually a whole art that our Swiss friends have passed down from generation to generation, as Eddy Baillifard, chef of the restaurant, reminds us. Raclett’Housewho lent his gestures to the robot:
“Teaching a living being is easy. I taught my son and stepson how to scrape, but when you have an arm that weighs 30 pounds in your hands, it’s yet another challenge. »
The master cheesemaker really taught the robot gestures by manipulating it. Indeed, the machine was not programmed to make the raclette, but records the trajectory of the scraper during manipulation, while using its various sensors to measure the force applied to the cheese. Emmanuel Pignat, doctoral student in the robotics group of the research institute ofIdiapprecise :
“By showing at different levels of melting, the robot learns that it is important to keep constant pressure, no matter how high the cheese is. It extracts the characteristics of position, force and speed necessary for an adaptive and elegant gesture. »
At each demonstration, the master cheesemaker performs slightly different gestures that are recorded by the robot’s learning algorithms. The latter incorporates these nuances to recreate the gesture in a new situation.
Many fields of application
This innovation, which can make you smile, shows the potential of artificial intelligence and machine learning to develop real collaboration between humans and machines. We can then imagine all kinds of applications in industry, services, but also to design assistance robots in order to relieve professionals of their dangerous, uncomfortable or repetitive tasks.
The objective is therefore to transmit new skills to the robot in an intuitive way, thanks to user-friendly interfaces and algorithms that will allow robots to learn by imitation. All without computer programming. Learning can then be done by observation, problem matching, or kinesthetic teaching.
By presenting their “Roboclette”, the engineers succeeded in attracting the attention of the public to highlight a promising technology that goes far beyond the convivial meal that raclette enthusiasts like to share.
JDG
