SculptBot: Pre-Trained Models for 3D Deformable Object Manipulation

TL;DR

This paper introduces SculptBot, a system that uses pre-trained point cloud models and a novel action sampling algorithm to enable robotic manipulation and sculpting of deformable objects like clay in real-world settings.

Contribution

It presents a new approach combining point cloud-based state representation, a pre-trained Transformer for dynamics prediction, and a geometry-aware action sampling method for deformable object manipulation.

Findings

Successfully models clay deformation dynamics.

Able to create simple shapes through robotic sculpting.

Operates entirely in real-world environments.

Abstract

Deformable object manipulation presents a unique set of challenges in robotic manipulation by exhibiting high degrees of freedom and severe self-occlusion. State representation for materials that exhibit plastic behavior, like modeling clay or bread dough, is also difficult because they permanently deform under stress and are constantly changing shape. In this work, we investigate each of these challenges using the task of robotic sculpting with a parallel gripper. We propose a system that uses point clouds as the state representation and leverages pre-trained point cloud reconstruction Transformer to learn a latent dynamics model to predict material deformations given a grasp action. We design a novel action sampling algorithm that reasons about geometrical differences between point clouds to further improve the efficiency of model-based planners. All data and experiments are conducted entirely in the real world. Our experiments show the proposed system is able to successfully capture the dynamics of clay, and is able to create a variety of simple shapes.