Description

The approach of the thesis will be to define action primitives that rely on robust low-level control to enable dynamics, and high-level planning to combine such actions, for a novel paradigm of Task-And-Motion planning considering dynamics (DTAMP). Exploiting the group’s previous results, we can also explore methods to reason in the C-space, explainability, perception of complex objects, etc.

The work is under the scope of the following projects:

• CHLOE-Map: Geometric Coordinates to Navigate the Configuration Space of Cloth-Like Objects for Robotic Manipulation (web)
• ROMANDIC: RObot MANipulation of Deformables through dynamIC actions (web)