Manipulation of Granular Materials by Learning Particle Interactions

TL;DR

This paper introduces a graph neural network-based approach to model and manipulate granular materials, enabling precise control of their configurations through learned particle interactions in both simulated and real environments.

Contribution

It presents a novel graph-based modeling and planning framework for granular materials using GNNs and Wasserstein distance minimization, addressing limitations of prior simplified dynamics models.

Findings

Successful manipulation of granular materials into desired configurations

Effective modeling of particle interactions with GNNs

Demonstrated applicability in both simulation and real-world scenarios

Abstract

Manipulation of granular materials such as sand or rice remains an unsolved problem due to challenges such as the difficulty of defining their configuration or modeling the materials and their particles interactions. Current approaches tend to simplify the material dynamics and omit the interactions between the particles. In this paper, we propose to use a graph-based representation to model the interaction dynamics of the material and rigid bodies manipulating it. This allows the planning of manipulation trajectories to reach a desired configuration of the material. We use a graph neural network (GNN) to model the particle interactions via message-passing. To plan manipulation trajectories, we propose to minimise the Wasserstein distance between a predicted distribution of granular particles and their desired configuration. We demonstrate that the proposed method is able to pour granular materials into the desired configuration both in simulated and real scenarios.