Principal Investigator: Carme Torras
Email: torras@iri.upc.edu
Web:https://www.iri.upc.edu/research/perception
Brief Theme Description:
Textile objects pervade assistive and domestic environments, and their versatile manipulation by robots opens up a wide range of possibilities. These advancements could enhance the autonomy of elderly and disabled individuals and streamline tasks in housekeeping and hospital logistics. However, while efficient procedures exist for robotic handling of rigid objects and virtual rendering of deformable objects, real-world dynamic cloth manipulation remains challenging. The vast number of degrees of freedom involved in non-rigid deformations introduces uncertainties in perception and action outcomes, making this task elusive.
The ERC Advanced Grant project CLOTHILDE (https://clothilde.iri.upc.edu/) made substantial progress in overcoming these challenges by combining powerful tools from topology and ML. Key achievements include skill learning from demonstration, task-oriented cloth representation, model-based predictive control, and innovative gripper design.
Building on these advancements, the proposed fellowship aims to extend the system’s capabilities to enable dynamic cloth manipulation. To achieve this, it will address one or more of the following challenges: fast cloth perception, compact cloth state representation, physically realistic and efficient simulation of cloth dynamics, new closed-loop control strategies, learning of robust grasping and manipulation skills, possibly through virtual reality, and human-robot collaborative applications such as dressing, bed making or cloth folding and unfolding.
Available Infrastructures: Fully equipped Robot Perception and Manipulation Laboratory, see https://www.iri.upc.edu/research/perception#facilities.
Possible Secondments: Several possibilities, preferably within Europe.
Keywords: Cloth Manipulation; Robotics; Physical Simulation; Cloth Dynamics; Grasping; Manipulation Mechanics.
