NeRP: Neural Rearrangement Planning for Unknown Objects

TL;DR

NeRP is a deep learning approach for multi-step object rearrangement that generalizes from simulation to real-world unseen objects, outperforming baselines in efficiency and planning time.

Contribution

We introduce NeRP, a neural network-based method for rearranging unseen objects that generalizes from simulation to real-world applications.

Findings

NeRP outperforms naive and model-based baselines.

NeRP arranges objects in fewer steps and with less planning time.

NeRP successfully operates on real-world rearrangement tasks.

Abstract

Robots will be expected to manipulate a wide variety of objects in complex and arbitrary ways as they become more widely used in human environments. As such, the rearrangement of objects has been noted to be an important benchmark for AI capabilities in recent years. We propose NeRP (Neural Rearrangement Planning), a deep learning based approach for multi-step neural object rearrangement planning which works with never-before-seen objects, that is trained on simulation data, and generalizes to the real world. We compare NeRP to several naive and model-based baselines, demonstrating that our approach is measurably better and can efficiently arrange unseen objects in fewer steps and with less planning time. Finally, we demonstrate it on several challenging rearrangement problems in the real world.