F&P: Embodied AI Through Morphological Adaptation in Robots

This week's Food & Paper will be given by Kyrre Glette from RITMO on embodied AI in robots.

Kyrre Glette


This presentation will present the process and results from our paper "Real-world embodied AI through a morphologically adaptive quadruped robot" (Nature Machine Intelligence). The paper shows that it can be helpful for robots to adapt their body when encountering new terrains. To demonstrate this, we have developed the DyRET quadruped robot that can automatically adjust its legs' length.


Kyrre Glette is an Associate Professor at the Robotics and Intelligent Systems group (ROBIN), Dept. of Informatics, and the RITMO Centre for Interdisciplinary Studies in Rhythm, Time and Motion. His research interests are in algorithms for artificial intelligence and adaptivity and how this can be applied in robotic systems. In particular, he studies evolutionary algorithms and computational self-awareness, both in collectives and single robot locomotion.

Published Sep. 28, 2021 2:19 PM - Last modified Dec. 9, 2021 3:46 PM