Towards synthesizing grasps for 3D deformable objects with physics-based simulation

TL;DR

This paper introduces a deep learning approach for grasping deformable objects by leveraging physics-based simulation to generate synthetic training data, enabling the model to adapt grasps based on object stiffness.

Contribution

It presents the first deep learning method that generates stiffness-dependent grasps for deformable objects using purely synthetic data from physics-based simulators.

Findings

Improved grasp ranking and success rate in simulation.

Network adapts grasps based on object stiffness.

Validated on synthetic data and ongoing physical robot tests.

Abstract

Grasping deformable objects is not well researched due to the complexity in modelling and simulating the dynamic behavior of such objects. However, with the rapid development of physics-based simulators that support soft bodies, the research gap between rigid and deformable objects is getting smaller. To leverage the capability of such simulators and to challenge the assumption that has guided robotic grasping research so far, i.e., object rigidity, we proposed a deep-learning based approach that generates stiffness-dependent grasps. Our network is trained on purely synthetic data generated from a physics-based simulator. The same simulator is also used to evaluate the trained network. The results show improvement in terms of grasp ranking and grasp success rate. Furthermore, our network can adapt the grasps based on the stiffness. We are currently validating the proposed approach on a larger test dataset in simulation and on a physical robot.