A Differentiable Recipe for Learning Visual Non-Prehensile Planar Manipulation

Abstract

Specifying tasks with videos is a powerful technique towards acquiring novel and general robot skills. However, reasoning over mechanics and dexterous interactions can make it challenging to scale learning contact-rich manipulation. In this work, we focus on the problem of visual non-prehensile planar manipulation: given a video of an object in planar motion, find contact-aware robot actions that reproduce the same object motion. We propose a novel architecture, Differentiable Learning for Manipulation (DLM), that combines video decoding neural models with priors from contact mechanics by leveraging differentiable optimization and finite difference based simulation. Through extensive simulated experiments, we investigate the interplay between traditional model-based techniques and modern deep learning approaches. We find that our modular and fully differentiable architecture performs better than learning-only methods on unseen objects and motions. https://github.com/baceituno/dlm.

Imitation Learning for Nonprehensile Manipulation Through Self-Supervised Learning Considering Motion Speed

“…In [8], Planning for sliding and moving an object considering friction and for carrying an object with it considering gravity and inertia are very difficult. Fotheremore, these methods assume the use of known object shapes, postures, material, and desired trajectories, which is computationally expensive and difficult to adapt to new objects and environments [9].…”