D-Grasp: Physically Plausible Dynamic Grasp Synthesis for Hand-Object Interactions

TL;DR

This paper presents D-Grasp, a reinforcement learning-based method using physics simulation to generate realistic, dynamic hand-object interaction sequences for moving objects to target poses, addressing complex articulation and physical interaction challenges.

Contribution

The paper introduces a hierarchical reinforcement learning framework with physics simulation for dynamic grasp synthesis, enabling realistic and stable hand-object interaction sequences.

Findings

Generates stable grasps and diverse motion sequences.

Can correct imperfect labels to produce plausible interactions.

Achieves human-like hand-object interactions in simulation.

Abstract

We introduce the dynamic grasp synthesis task: given an object with a known 6D pose and a grasp reference, our goal is to generate motions that move the object to a target 6D pose. This is challenging, because it requires reasoning about the complex articulation of the human hand and the intricate physical interaction with the object. We propose a novel method that frames this problem in the reinforcement learning framework and leverages a physics simulation, both to learn and to evaluate such dynamic interactions. A hierarchical approach decomposes the task into low-level grasping and high-level motion synthesis. It can be used to generate novel hand sequences that approach, grasp, and move an object to a desired location, while retaining human-likeness. We show that our approach leads to stable grasps and generates a wide range of motions. Furthermore, even imperfect labels can be corrected by our method to generate dynamic interaction sequences.