Digital twins for fluid sloshing problems

Teaching robots to manipulate liquids (to help elderly people, for instance) is a challenge. We are used to recognize different fluids at first sight, but this educated intuition is difficult to transmit to robots.

Our last job has been just published in the Public Library of Science (PLOS)ONE journal (Open Access):

We present a #digitaltwin for fluid sloshing that, through computer vision and making an extensive use of simulation in the background, is able to learn the particular fluid and its behavior—even if the robot has not been exposed to it before—and predict its dynamics in the near future.

A great job by BEATRIZ MOYA, David González, Iciar Alfaro, and Francisco Chinesta, under the financial support from ESI Group, which is gratefully acknowledged.