Learning control strategy in soft robotics through a set of configuration spaces

Ménager, Etienne; Duriez, Christian

Document type :

Autre communication scientifique (congrès sans actes - poster - séminaire...): Communication dans un congrès avec actes

Title :

Learning control strategy in soft robotics through a set of configuration spaces

Author(s) :

Ménager, Etienne [Auteur]
Deformable Robots Simulation Team [DEFROST ]
Duriez, Christian [Auteur]

Inria Lille - Nord Europe
Université de Lille
Deformable Robots Simulation Team [DEFROST ]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]

Conference title :

IEEE International Conference on Soft Robotics (RoboSoft)

Conference organizers(s) :

IEEE

City :

San Diego (CA)

Country :

Etats-Unis d'Amérique

Start date of the conference :

2024-04

Publication date :

2024-04

English keyword(s) :

Modelling Control and Learning for Soft Robots.
Soft Robot Applications

HAL domain(s) :

Informatique [cs]/Robotique [cs.RO]

English abstract : [en]

The ability of a soft robot to perform specific tasks is determined by its contact configuration, and transitioning between configurations is often necessary to reach a desired position or manipulate an object. Based on ...
Show more >The ability of a soft robot to perform specific tasks is determined by its contact configuration, and transitioning between configurations is often necessary to reach a desired position or manipulate an object. Based on this observation, we propose a method for controlling soft robots that involves defining a graph of configuration spaces. Different agents, whether learned or not (convex optimization, expert trajectory, and collision detection), use the structure of the graph to solve the desired task. The graph and the agents are part of the prior knowledge that is intuitively integrated into the learning process. They are used to combine different optimization methods, improve sample efficiency, and provide interpretability. We construct the graph based on the contact configurations and demonstrate its effectiveness through two scenarios, a deformable beam in contact with its environment and a soft manipulator, where it outperforms the baseline in terms of stability, learning speed, and interpretability.Show less >

Language :

Anglais

Peer reviewed article :

Oui

Audience :

Internationale

Popular science :

Non

ANR Project :

Infrastructure technologique pour la recherche d'excellence en robotique

Collections :