DefGraspSim: Physics-Based Simulation of Grasp Outcomes for 3D Deformable Objects

Robotic grasping of 3D deformable objects (e.g., fruits/vegetables, internal organs, bottles/boxes) is critical for real-world applications such as food processing, robotic surgery, and household automation. However, developing grasp strategies uniquely challenging. Unlike rigid objects, have infinite degrees freedom require field quantities deformation, stress) to fully define their state. As these are not easily accessible in the real world, we propose studying interaction with through physics-based simulation. such, simulate grasps on a wide range using GPU-based implementation corotational finite element method (FEM). To facilitate future research, open-source our simulated dataset (34 1e5 Pa elasticity range, 6800 evaluations, 1.1M measurements), well code repository that allows researchers run full FEM-based evaluation pipeline arbitrary object models choice. Finally, demonstrate good correspondence between outcomes counterparts.