Description
Robotic manipulation of cloth presents a complex challenge due to the infinite-dimensional shape-state space of textiles. This complexity makes accurate state estimation a daunting task. To address this issue, we introduce the concept of dGLI Cloth Coordinates—a finite, low-dimensional representation of cloth states. This novel approach enables us to effectively distinguish among a wide range of folded states, paving the way for efficient learning methods in cloth manipulation planning and control.
