Grasp'D: Differentiable Contact-rich Grasp Synthesis for Multi-fingered Hands

Dylan Turpin^1,2,3, Liquan Wang^1,2,3, Eric Heiden³, Yun-Chun Chen^1,2, Miles Macklin³, Stavros Tsogkas⁴, Sven Dickinson^1,2,4, Animesh Garg^1,2,3
1University of Toronto ²Vector Institute, ³Nvidia, ⁴Samsung

European Conference on Computer Vision (ECCV) 2022

Abstract. The study of hand-object interaction requires generating viable grasp poses for high-dimensional multi-finger models, often relying on analytic grasp synthesis which tends to produce brittle and unnatural results. This paper presents Grasp’D, an approach for grasp synthesis with a differentiable contact simulation from both known models as well as visual inputs. We use gradient-based methods as an alternative to sampling-based grasp synthesis, which fails without simplifying assumptions, such as pre-specified contact locations and eigengrasps. Such assumptions limit grasp discovery and, in particular, exclude high-contact power grasps. In contrast, our simulation-based approach allows for stable, efficient, physically realistic, high-contact grasp synthesis, even for gripper morphologies with high-degrees of freedom. We identify and address challenges in making grasp simulation amenable to gradient-based optimization, such as non-smooth object surface geometry, contact sparsity, and a rugged optimization landscape. Grasp’D compares favorably to analytic grasp synthesis on human and robotic hand models, and resultant grasps achieve over 4× denser contact, leading to significantly higher grasp stability.

• Code
• Paper
• arXiv
• Video walkthrough